Tracking changes to AI fintech vendors

Fenergo added a twice daily option for Ongoing Screening. An enabled client keeps its existing daily run and gets a second run 12 hours later, so a sanctions list update issued after the first run is picked up about 12 hours later instead of on the next working day. Fenergo turns it on at a client's request, and clients who do not need it keep the once a day run.

Our read

Sanctions teams can shorten the gap between a new designation and a match against existing clients. There is no self service control and no custom run times, so ask your Fenergo representative to confirm availability before relying on it.

Broadridge's CQG trading platform now offers Kalshi event contracts through its front ends and APIs, next to futures, options and equities. Brokers and futures commission merchants can give clients direct trading access and market data. Traders can track event contract positions with CQG's risk management and portfolio monitoring tools. The launch serves client demand in the United States and EMEA, with access subject to eligibility and jurisdictional rules.

Our read

Brokers already on CQG can add prediction markets without a separate venue connection. Access for the more than 1,700 buy side clients on the NYFIX network is described as coming rather than live, so NYFIX users should confirm timing before planning around it.

Dynamic Planner released an agentic upgrade to its Digital Advice Platform at its Transforming Advice: Revealed event in London. Connectors to Microsoft 365 and Zoom are now live, and firms can push all fact find data back to Intelligent Office. A new license lets advice firms build their own multi channel client journeys on Dynamic Planner's data, risk models and insights.

Our read

Firms on Intelligent Office should confirm which fact find fields now write back and how conflicts with existing records are handled. The release does not say what the Microsoft 365 and Zoom connectors do or how the new license is priced, so ask for both.

GBG added a face uniqueness check to GBG Go that flags duplicate sign ups at registration, so one face holds one account. The same face check can run as a step up after a device change, unusual withdrawal or account recovery request, and it returns a reusable face credential rather than a stored biometric image. GBG says both capabilities are available through GBG Go from October 2026.

Our read

Fraud teams can stop people opening several accounts to collect sign up bonuses before a promotion pays out. The face credential comes from Realeyes technology, so check where it is held and how your biometric consent terms cover it.

Insurity released Dorado, led by its Ratio Content & Rating Engine, an API first engine that turns machine readable bureau rates, rules, forms and statistical codes into production insurance logic and, Insurity says, can rate thousands of vehicles in under a minute. The release also adds AI powered insights, natural language analysis and expanded risk data for underwriters, new SpatialKey capabilities that make geospatial analysis easier for business users, and changes across policy administration, claims and financial operations.

Our read

Carriers writing small and complex commercial lines can test whether bureau content updates reach production faster on one rating engine. The release gives no accuracy figures for the AI analysis, so underwriting leaders should compare its observations with their existing accumulation and catastrophe views before relying on them.

Middesk expanded International Business Verification to more than 65 countries, more than doubling its coverage and reaching Latin America for the first time. New markets in Latin America and the Caribbean, the Middle East and Africa return registry data through the same endpoint, API parameters, batch CSV templates and dashboard used for US checks. The release also adds official registry extracts and formation documents in 18 countries.

Our read

Banks and payments firms onboarding foreign businesses can cover more jurisdictions without a second KYB vendor or a new integration. Middesk does not publish the country list, so ask which markets are covered and which return documents before relying on it.

Moveo AI added a Debts page to its payment recovery product that lists every debt in an account in one place, including debts not linked to a contact. Users can search by external ID or contact name, and filter by status, promise to pay, compliance flags or linked contact. Opening a debt shows its balance, compliance flags and linked contacts, and debts can be deleted from the page.

Our read

Collections and lending teams running AI agents on borrower conversations get one view of the compliance flags across the portfolio, such as disputes. Check who on the team can delete a flagged debt, since that removes the record of what the agent captured.

Napier AI integrated its Transaction Screening product with LSEG's World-Check On Demand API after completing LSEG's technology validation. Payments can now be screened against World-Check sanctions, politically exposed person, adverse media and enforcement data within minutes of LSEG analysts publishing it. Firms set which transactions need the check by payment rail and customer risk type, and can screen only against the sources that fit their risk appetite.

Our read

Banks and payment firms that license World-Check can apply new list changes to live payments sooner, with source level provenance on each match. Test whether rail and risk type rules actually cut false positives on your payment mix before switching from file based list loads.

Tyfone

Oct 29, 2026

ImpactMedium

VerificationPartially Verified

Tyfone added a prebuilt integration that lets credit unions and community banks on its digital banking platform switch on Pocketnest's financial wellness guidance without a custom integration project. Members get customized recommendations, goal tracking and to do lists covering budgeting, debt, savings, retirement and estate planning inside online banking. ELGA Credit Union, with more than 117,000 members and $2.4 billion in assets, is the first to launch it.

Our read

Tyfone clients can add guidance to digital banking by enabling an existing connector rather than funding a build. No technical documentation is public, so confirm what member data flows to Pocketnest and who controls the recommendations members see.

Fingerprint

Oct 28, 2026

ImpactMedium

VerificationVerified

Fingerprint's Bot Detection Smart Signal now labels traffic inferred to be Instinct, a personal AI agent from Spear Street Technology that users direct by text message. Responses return bot_type "instinct" with the provider, an unknown identity flag and a high, medium or low confidence value, through the Server API, the MCP server and dashboard filters. Detection is rolling out to all accounts over a week, at no extra cost for Pro and Enterprise customers using Bot Detection.

Our read

Instinct logs in with a customer's stored credentials and relays MFA codes, so it can pass for the account holder. Banks and lenders can use the label to step up verification on account changes or rate limit sessions rather than blocking a real customer's agent outright.

FintechOS

Oct 28, 2026

ImpactMedium

VerificationVerified

FintechOS released Backoffice Insurance 8.1.1 with an API that lets authorized underwriters override automated underwriting results on offer cards held for referral. Each override logs a mandatory business justification and the acting user, keeps the original rule results, and creates a new card version. A companion endpoint returns rule level underwriting outcomes across insurance and banking products and hides internal formula steps by default.

Our read

Insurers get a documented human review path for referred quotes, with a record of who changed what and why. Overrides apply only to referred cards, so declined cards still need a separate route if your process allows exceptions.

K1x

Oct 28, 2026

ImpactMedium

VerificationVerified

K1x added an Investor Portal to K1 Creator in its 10/28/2026 release, so issuers can deliver K-1 packages to investors through a secure portal. The release also made Investor Management available, a reusable investor library kept at the tenant level and separate from entity specific associations, along with Entity Management for associating investors to entities. The same release completed the CCH Tax Year 2025 integration for supported jurisdictions.

Our read

Funds and fund administrators that issue partner K-1s can keep investor records and package delivery in the same tool that builds the forms. Check how the portal handles investor access and notifications before retiring a separate delivery provider.

MyComplianceOffice

Oct 28, 2026

ImpactMedium

VerificationVerified

MyComplianceOffice (MCO) now takes in business messages from WhatsApp, WeChat, LINE, Telegram and SMS that firms enable and record through LeapXpert. The captured messages land in MCO's compliance environment for archiving and review. Compliance teams then manage that oversight in the same program as employee trading, conflicts, disclosures and investigations.

Our read

Firms under off channel recordkeeping pressure can sanction consumer apps without standing up a separate archive. Capture depends on LeapXpert, so confirm licensing, setup and which channels each vendor covers before relying on it for books and records.

Nasdaq Verafin

Oct 28, 2026

ImpactHigh

VerificationPartially Verified

Nasdaq Verafin said that over the past month it began rolling out automatic alert dispositioning for its Agentic AI Workforce. Financial institutions can now configure the workforce's agentic analysts to clear appropriate alerts on their own, without an analyst signing off, or to escalate them for further human review.

Our read

The agentic analysts now close alerts without a person instead of only recommending an outcome. Ask which agents and alert types your rollout covers, and check that the record each automated closure leaves will satisfy examiners and model risk reviewers.

Proof

Oct 28, 2026

ImpactMedium

VerificationVerified

Proof added Authentication Classes to Proof Digital ID, so a business can set the authentication level each transaction type requires. Notarization and identity verification still use full biometric authentication, while payments and AI agent actions can be authorized with lighter methods such as SMS OTP. Every action traces back to the user's Digital ID enrolled at IAL2, and Proof enforces the set level and keeps an auditable record of each authorization.

Our read

Banks and payment firms that accept Proof Digital ID can lower friction on routine payments and keep biometric checks for higher stakes actions. Before mapping agent initiated payments to an SMS OTP class, confirm that it fits your own authentication risk policy.

Shufti

Oct 28, 2026

ImpactHigh

VerificationVerified

Shufti launched Know Your Agent (KYA), which verifies AI agents that open accounts, place orders or move funds for customers. Each agent links to a person or business verified through Shufti KYC or KYB, and gets a revocable, time bound credential and a trust score. Every request is checked against a mandate of permitted actions, limits and expiry, and requests outside it or above a high risk threshold trigger a selfie liveness check by the authorizing person.

Our read

Banks get a way to tie each agent payment or account action to an accountable, verified person, with an audit trail of requests, decisions and revocations. Shufti's developer revision history lists no KYA documentation yet, so confirm the integration and how mandates are set in a demo.

Verisk

Oct 28, 2026

ImpactMedium

VerificationVerified

Verisk added Amazon Quick and Microsoft Copilot integrations that give eligible insurers natural language access to ISO Loss Costs and Premium Audit Advisory Service (PAAS) content. Its existing Verisk Underwriting Intelligence in Claude now covers PAAS resources alongside ISO Loss Costs. Verisk also added a Benchmark connector in Claude for permissioned access to property level hail and wind history.

Our read

Carriers that use Amazon Quick or Microsoft Copilot can now query ISO and PAAS content without moving users to a separate tool. Verisk says existing entitlements, permissions and governance controls still apply, so confirm which licenses cover the new Benchmark weather connector.

Zocks

Oct 28, 2026

ImpactHigh

VerificationVerified

Zocks launched Account Opening, which fills and validates new account applications from intake meeting data, connected CRM records and processed documents before they are submitted. Advisors who custody with Schwab can now send a validated application through Schwab Advisor Center's digital onboarding workflow, with clients authorizing through Schwab Alliance credentials. Coverage includes Schwab One brokerage registrations and traditional, Roth, rollover and inherited IRAs, with SEP and SIMPLE IRAs listed as coming soon.

Our read

Conversation data now flows into a custodial submission rather than notes alone, so supervision teams should review how validation works and where approval sits. Schwab is the only custodian with direct submission so far; other firms rely on links to their own workflow systems.

Backbase

Oct 27, 2026

ImpactHigh

VerificationVerified

Backbase released Delivery Factory, generally available in all markets, where a bank's delivery team works with specialist AI agents from agreed requirements to working code. Business Analyst, Solution Architect, Backend, Mobile and Web Agents share one project context built from scoping documents, architecture specifications and Backbase product knowledge, and produce specifications, Jira work items and implementation plans. Backbase reports teams using it delivered twice as fast as traditional teams across more than 10 implementations.

Our read

Engineers review every code change before work moves forward, so the bank's own team stays accountable for what ships. The Migration and Upgrade Agents introduced alongside it have no stated availability, so confirm their status before planning a migration or upgrade around them.

Conquest Planning

Oct 27, 2026

ImpactMedium

VerificationVerified

Conquest released an MCP connector that brings its Strategic Advice Manager (SAM) planning engine into Claude for Financial Advisors, and it is available now in Anthropic's connector directory. Advisors can pull client plan details and summaries, check goal progress, surface planning opportunities, create client files, and update plans or apply strategies from the chat. Conquest says results come from SAM's deterministic calculation engine, which ranks and verifies each strategy.

Our read

Because advisors can change live plans from the assistant, test how confirmation steps and the audit trail behave before rollout. Conquest says its regional hosting in the US, Canada and the UK lets firms control where their data resides.

FE fundinfo

Oct 27, 2026

ImpactMedium

VerificationVerified

FE fundinfo added Monte Carlo stress testing to the UK gross cashflow modeler in FE CashCalc, its cashflow planning software for advisers. Advisers can run 10,000 simulations across one pot or any combination of pots, using the contributions, withdrawals and charges already in the plan, with no rekeying into the net modeler. Results show 25th, 50th and 75th percentile yearly breakdowns and export as a PDF report.

Our read

Advice firms can now stress test plans in the modeler that calculates the tax due, and keep one report of pots, parameters and charges on the client file. It is included in existing subscriptions, and the Irish gross modeler is expected by the end of 2026.

Jump

Oct 27, 2026

ImpactMedium

VerificationVerified

Jump released eight prebuilt Claude skills that work from an advisor's Jump meeting transcripts, notes, documents, contacts and tasks. They include Client Commitments, Decision Log, Meeting Recap, Recap Email and Document Lookup, and run by slash command or plain language. Jump's existing connector, which gives Claude read only access, now installs as part of Anthropic's Claude for Financial Advisors plugin.

Our read

Firms that allow Claude can let advisors use meeting history without copying notes, while Jump's recording, retention, consent and access settings still apply. Jump points out that Anthropic recommends Enterprise plans for registered investment advisers because they include the audit logs that support recordkeeping.

Nitrogen

Oct 27, 2026

ImpactHigh

VerificationVerified

Nitrogen began rolling out Nucleus Account Opening, which lets advisors start, submit and track new Charles Schwab accounts inside Nitrogen using Schwab's new account opening APIs. Nucleus, its agentic AI engine, prefills details from existing client and household data, and a pending account appears in the client's portfolio. When Schwab opens the account, the account number flows back and Nitrogen reconciles it automatically.

Our read

Advisors review, confirm and submit each request, so the agent prepares the paperwork but does not open accounts on its own. Schwab is the only custodian so far, and the rollout covers eligible customers with a connected Schwab integration.

Scorechain

Oct 27, 2026

ImpactMedium

VerificationVerified

Scorechain added support for Arc, Circle's Layer 1 network for stablecoin payments and tokenized assets, as the 26th blockchain on its platform. Compliance teams can screen Arc addresses, apply their alert rules to Arc deposits and withdrawals, track USDC on Arc and trace funds between Arc and other chains. Coverage is live across every product and the Blockchain Analytics API for all clients, with no setup needed.

Our read

Banks, custodians and payment firms touching USDC or tokenized funds on Arc can screen there under their current risk settings. Arc settles in under a second, so screening has to run before a payment is sent, which makes the API path the one to test.

Taktile

Oct 27, 2026

ImpactMedium

VerificationVerified

Taktile launched Insights, an AI analytics feature inside its decision platform, after a customer beta. Operations leaders, P&L owners and analysts at banks and insurers ask plain language questions about their decision data and get answers and charts back. They can pin those charts into their own dashboards and a shared control tower, then drill from product, segment or geography level down to individual flows and decisions.

Our read

Credit, fraud and claims owners can trace a shift in approval rates or other KPIs to the flow behind it without waiting on a separate BI team. Before using it for board or monitoring reports, check how generated charts can be traced back to the underlying decision data.

Zeplyn

Oct 27, 2026

ImpactMedium

VerificationVerified

Zeplyn introduced Zeplyn Workflows, which turns a wealth firm's operating playbooks into multi step processes that run across systems and departments, starting with advisor transitions, client onboarding and account opening. AI agents decide next steps within set boundaries, while fixed rules validate data and update systems. Each workflow pauses at firm defined approval points for a designated person and keeps an owner, status and activity history.

Our read

This moves Zeplyn from advisor notes into middle and back office work, so operations and compliance teams should map which steps the agents take on their own. Test where approval points sit for account opening, and whether the activity history meets supervision and books and records needs.

AI Rudder

Oct 26, 2026

ImpactLow

VerificationVerified

AI Rudder moved the HubSpot connector for its CALL-E voice agent onto Calls API v2. The HubSpot Contact card and workflow action now recover from timeouts, and a replayed request with the same key does not place a duplicate call. AI Rudder also published a v2 n8n workflow template and CALL-E plugin 0.2.2 on the Dify Marketplace, with a separate v2 Dify template on GitHub.

Our read

Teams that trigger customer calls from HubSpot or workflow tools now have a supported path off v1 Calls, which retire on December 31, 2026. Existing n8n workflows need an explicit upgrade, and the v2 Dify template is still under Marketplace review.

Bits Technology

Oct 26, 2026

ImpactMedium

VerificationVerified

Bits Technology added goAML Suspicious Activity Report and Suspicious Transaction Report filing for Sweden and Finland to its Reports hub. A guided form covers report info, attachments, indicators, parties, goods and services, and a preview, and STR drafts add transactions with a CSV import. Individuals, organizations and bank accounts can be prefilled from entities already in Bits, and one click downloads the goAML XML file for upload to the financial intelligence unit portal.

Our read

Filing from the same records used in the investigation cuts retyping and identifier mismatches, and leaves an archive of what was reported and when. Coverage is Sweden and Finland only, and the downloaded XML still has to be uploaded to the FIU portal.

Codat

Oct 26, 2026

ImpactLow

VerificationVerified

Codat rebuilt Intelligent Upload in Link, the path for businesses whose accounting software has no direct integration, into one guided screen. Customers upload CSV or XLSX files for Vendors, Bills, Payments and optionally Expenses covering the last 12 months, can replace a file before submitting, and resend only missing or unusable files if the connection drops. Lenders can add a support email and a scheduling link in Link settings or through the Link SDK.

Our read

Banks using Spend Insights should see more complete uploads from clients whose ERP has no connector. The new journey applies to every customer automatically and is offered in English and Canadian French only, so check coverage for other client languages.

DigiFi

Oct 26, 2026

ImpactHigh

VerificationVerified

DigiFi added four Jack Henry SilverLake core banking integrations to its Integration Marketplace: Search Customer, Create Customer, Create Account and Create Collateral. Lenders add them to a loan product through Data Integrations, so the origination workflow can look up an existing SilverLake customer and create the customer, loan account and collateral records in SilverLake.

Our read

Banks on SilverLake can create core customer, loan account and collateral records from DigiFi instead of rekeying them. Lenders should test field mapping and duplicate customer handling, since the release covers only these four endpoints.

Earnix

Oct 26, 2026

ImpactLow

VerificationVerified

Earnix released its Q4 2026 pricing and rating enhancements, led by a new Online Testing Hub where analysts generate and run API requests in a staging environment to check pricing and underwriting rules before production. Package and multistate rating now prices several products and states through a single API request. Automated base rate adjustments reach a target overall rate change without recalculating base rates in spreadsheets.

Our read

Carriers with multistate or packaged programs can price them in one request, which Earnix says simplifies integration, so it is worth testing against current quoting systems. Pricing teams should confirm how staging test results feed their change approval and filing evidence.

HyperVerge

Oct 26, 2026

ImpactLow

VerificationPartially Verified

HyperVerge added emulator and simulator detection to WebCore mode in HyperKYC Flutter SDK 3.6.0, on both Android and iOS. The check reports an isEmulator signal and is enforced through the exitIfEmulator workflow flag, which ends the session with error 142. The release bundles native Android HyperKYC 3.6.0, iOS HyperKYC 1.12.0 and WebCore 11.6.1.

Our read

Lenders and banks that run HyperVerge's web based capture can now stop onboarding sessions launched from emulators, a common route for scripted account opening. The named flag and error code give fraud teams a loggable reason for each rejection, so check how it feeds case review.

iDenfy

Oct 26, 2026

ImpactLow

VerificationVerified

iDenfy's Sprint 138 lets dashboard users edit or delete their own comments on KYB companies, AML v3 checks and monitoring, and risk assessments. Edited comments are marked (edited), deleted ones drop out of the dashboard, PDF reports and exports, and every change is recorded in the audit trail. Deactivated and deleted team members now keep their name, surname and email on their past reviews, comments, status changes and rules.

Our read

Compliance teams keep a named reviewer on every past decision after staff leave, which is what examiners and auditors ask for. Because deleted comments vanish from reports and exports, confirm the audit trail keeps the original text for record retention.

Incode Technologies

Oct 26, 2026

ImpactHigh

VerificationVerified

Incode achieved ISO/IEC 42001:2023 certification for its AI management system, issued by Sensiba LLP, an ANAB accredited certification body, after an independent audit. The certificate covers the AI systems behind the Incode Omni identity platform across design and development, training and evaluation, deployment and monitoring. It was issued in October 2026, is valid through October 2029, and is viewable in the Incode Trust Center.

Our read

Banks reviewing an identity verification vendor under third party and model risk programs now have an accredited AI governance audit to cite. Ask for the certificate scope to confirm which models and products it covers before relying on it in a review.

LendAPI

Oct 26, 2026

ImpactLow

VerificationPartially Verified

LendAPI v3.3.0 added Branches, which lets a lender with several physical locations give each one its own staff, applications and slice of reporting. Branch Associates see only the applications they handle, and Branch Managers see their whole branch and manage their own associates. Applications are tagged to a branch through a branch application link or the staff member who creates them, and the branch appears in exports, webhooks, contracts, email templates and the API.

Our read

Credit unions and dealer or retail lenders can limit which staff see which applications without building workarounds, and carry branch attribution into their own systems. Branches is off by default and is turned on one sub account at a time, so plan a staged rollout.

Narmi

Oct 26, 2026

ImpactMedium

VerificationVerified

Narmi launched a Microsoft 365 Copilot connector built on its Model Context Protocol server, connecting Copilot to Narmi's administrative banking platform. Bank and credit union staff can ask in natural language about customer account activity, transaction histories, line of credit utilization, lending applications and account application status. Access follows each institution's existing permissions, and the connection is read only.

Our read

Institutions already on Microsoft 365 get a staff route into Narmi data without adding a platform. Because the assistant cannot modify accounts or initiate transactions, the review centers on what data staff can pull into Copilot and how that fits existing access and retention policies.

Spektr

Oct 26, 2026

ImpactMedium

VerificationVerified

Spektr moved the prepayment check in Website Sentinel from AI based detection to a wordlist scan. A new Prepayment threshold in the node config sets how many distinct listed words must appear before the check flags, with a default of 1, and each hit records the word, the page URL and its position. The same release stores a screenshot of every scanned page as evidence, though screenshots do not yet appear in scan results.

Our read

Merchant onboarding teams get a prepayment flag that traces to a specific word on a specific page, which is easier to explain to an examiner. Spektr warns results can differ from earlier scans, so current users should review the threshold before relying on new flags.

AllMind

Oct 25, 2026

ImpactHigh

VerificationVerified

AllMind said it will retire Reports in early November and is helping teams move their report workflows to Agent Studio and Skills, including automations that use Reports. Financial Models and Screener are temporarily unavailable while AllMind works on them. The company says it plans to bring both back in a future update but gives no date.

Our read

Teams that run recurring research through Reports, Financial Models or Screener need a migration plan now. Test whether Skills with the new Outlines setting reproduce existing report templates, and confirm in writing which modules the contract still covers.

Glia

Oct 25, 2026

ImpactMedium

VerificationPartially Verified

Glia's virtual assistant at University of Kentucky Credit Union, named Kat, now handles about 58 to 60% of calls, according to the credit union's AVP of Member Experience in a CUbroadcast interview. He reported a 47% containment rate, compared with 5% on the legacy contact center system Glia replaced. He also cited a 13% increase in referrals after staff were reallocated, with longer handle times as staff spent more time on member issues.

Our read

This is a named US credit union with before and after figures, but they are self reported in an interview rather than a documented study. Ask Glia how containment is defined and measured before using the 47% figure in a business case.

Advocate Technologies

Oct 23, 2026

ImpactMedium

VerificationPartially Verified

Advocate released AIR in beta, an agent that builds insurance requirements through conversation instead of field by field entry. A user describes what a program requires, and AIR works against the organization's existing risk profiles to propose and build the requirements. AIR shipped in release v8.32.1 and is invite only while Advocate works with early users.

Our read

Commercial loan servicers that set insurance requirements by hand can test whether AIR drafts them correctly against their own risk profiles. Access is by invitation for now, so confirm entry before planning a program rollout around it.

Argos Identity

Oct 23, 2026

ImpactMedium

VerificationVerified

Argos Identity now lets ARGOS Omni users fix the match threshold on an AML screening engine for a workflow step, anywhere from 75 to 100 with a default of 85. Person searches can also fix date of birth matching, which defaults to same year. Fixed values override what the AI agent sends, and the values actually used are stored in the agent audit log and the analysis results.

Our read

Compliance teams can stop the agent from loosening AML match sensitivity case by case, and can show examiners the exact criteria applied to each screening. Integrations can read the fixed options from the workflow API to confirm what each step enforces.

Blend

Oct 23, 2026

ImpactMedium

VerificationVerified

Blend now lets lenders run Freddie Mac's Loan Product Advisor (LPA) automatically when a borrower submits a full application and the automatic credit pull completes, matching the existing Desktop Underwriter automatic run. It is opt in through Blend support and enabled separately for Conventional, VA and FHA purchase loans. Findings appear in AUS Tracking in the LO Toolkit, and when both systems run, the most recent successful result is selected.

Our read

Lenders that rely on LPA, or run both systems, get findings at submit without a loan officer step. Check the skip rules first: refinances, VA purchases with a backend DTI of 41% or higher, missing REO linking and incomplete credit pulls still need a manual run.

FICO

Oct 23, 2026

ImpactMedium

VerificationVerified

FICO said it will hold its wholesale mortgage price for the FICO Score at $10 through tri merge resellers for lenders and brokers enrolled in its Preferred Program. The $10 covers a bundle of Classic FICO Score and FICO Score 10T. FICO describes the price as flat year over year and says it is available to lenders who have committed to FICO through participating reseller partners.

Our read

Mortgage lenders get Classic FICO and FICO Score 10T at one wholesale price, which helps plan a dual score process. The release does not spell out what the Preferred Program commitment requires or what lenders outside it pay, so get those terms in writing.

Hebbia

Oct 23, 2026

ImpactMedium

VerificationVerified

Hebbia added a Databricks integration that lets users ask plain language questions of a firm's own Databricks data, such as portfolio metrics, positions and fund performance, alongside documents, filings and market data. Data stays in the firm's Databricks environment, and Hebbia reads it without copying it out. Each user signs in with their own Databricks account, and every answer shows its source.

Our read

Credit, private equity, public equity and insurance investment teams can test KPIs against covenant terms or holdings against research without a data team request. Confirm that per user sign in maps to your existing Databricks entitlements before rollout.

ibl.ai

Oct 23, 2026

ImpactLow

VerificationVerified

ibl.ai updated its Chrome extension, which runs an organization's own ibl.ai agents in a Chrome side panel where Cowork reads and acts on the open tab. Between 21 and 23 October the agent moved to real mouse and keyboard events sent through Chrome's DevTools protocol, models that read images began receiving a picture of the tab, and the agent can now click or type at a point on that picture. Version 1.1.9 is on the Chrome Web Store.

Our read

Agents can now work inside web systems that offer no API, with model calls going through the institution's own ibl.ai session. The default Approvals mode is Automatic, so risk teams should require Ask Me, which asks before a submit, payment, send or delete, for agents that touch production systems.

Kore.ai

Oct 23, 2026

ImpactMedium

VerificationVerified

Kore.ai's Artemis agent platform v1.8.2 lets builders list the exact session fields the model may write in a reasoning step, instead of only turning model writes on or off. An empty list blocks all model supplied updates in that step, while authored actions such as SET, GATHER extraction, hooks and tool result mappings work as before. Existing agents keep their current behavior until a step is changed.

Our read

Teams building servicing agents can stop the model from writing fields such as verification flags or amounts, leaving those to authored logic. Every Runtime instance must run v1.8.2 before an agent that uses field lists is published.

SS&C Technologies

Oct 23, 2026

ImpactHigh

VerificationVerified

SS&C Blue Prism added a batch of new WorkHQ workflow connectors, most of them for financial services systems, including Jack Henry, Fiserv, FIS Code Connect, BNY Mellon, Northern Trust, State Street Alpha, DTCC, Broadridge, Equifax, TransUnion, FICO, Plaid and Refinitiv World-Check One. Each connector can start a workflow on events and take actions in that system. The Jack Henry connector reacts to account, customer and transaction events on the jXchange core and can post transactions, transfer funds and issue EFT cards.

Our read

Banks can now have agentic workflows write to the core, not just read from it, so access scopes and audit fields need careful setup. SS&C says users must test community connectors end to end, and support for them is best efforts.

AMLBot

Oct 22, 2026

ImpactLow

VerificationVerified

AMLBot cut its lineup to two plans. Lite is now Personal and Pro+ is now Business, and Pro packages can no longer be bought in the account or bot. Personal is pay per check, with credits from $1 down to $0.10 each and a Low, Medium or High risk result. Business adds an exact 0% to 100% risk score, named entity connections, KYT and KYC API access and real time monitoring, priced on request.

Our read

AMLBot says both plans run the same risk scoring, so the gap is depth of output and API access rather than a different engine. Pro customers need to arrange their next top up through support, and Business remains quote only.

Bloomberg

Oct 22, 2026

ImpactMedium

VerificationVerified

Bloomberg added classifications to Crypto Exposure Analytics that map crypto and tokenized asset exposure to the Basel Committee's Group 1a, 1b, 2a and 2b categories. The look through covers exposure held inside funds, ETFs, equities, derivatives and corporate bonds, and each classification carries explainability fields so clients can review, validate and challenge the result. Data comes through Bloomberg Data License by SFTP, REST API or cloud delivery, and on the Terminal through the WSL CRYPTOEXPO and WSL CRYPTOETFS worksheets.

Our read

Banks implementing the Basel crypto standard can replace watchlists and spreadsheets with classifications that carry a stated rationale. Bloomberg says the classifications come from a rules based process, so test the explainability fields against your own holdings before using them for capital treatment.

Carta

Oct 22, 2026

ImpactHigh

VerificationVerified

Carta now lets fund teams take the final action on common fund workflows from Workhub in Claude. Workhub could already start workflows and track requests. Users can now release capital activity, approve expense payments, management fees and financial packages, reconcile cash transactions that need more information, update valuations and approve the schedule of investments.

Our read

Steps that release money now complete inside an AI assistant. Carta says they carry the same safeguards as its web app, so controllers should confirm that permissions, approval chains and the audit record match before staff use it.

Convr

Oct 22, 2026

ImpactMedium

VerificationVerified

Convr extended its AI underwriting workbench to the cyber insurance line. Underwriting teams can now ingest and validate cyber documents such as unstructured applications, security questionnaires and supplemental documentation in the same workbench they use for other commercial lines. The Risk Context Engine interprets cyber submissions against a domain specific knowledge base, so underwriters can trace what was extracted and why.

Our read

Carriers and MGAs writing cyber can retire a separate manual intake process for that line. The release names no customers and gives no accuracy figures, so test extraction on your own applications and questionnaires, which vary widely by carrier.

Jack Henry

Oct 22, 2026

ImpactMedium

VerificationPartially Verified

Jack Henry added API reference documentation for Digital Assets v1 Account Addresses to its Digital Core developer documentation. The new endpoints create, list and search account addresses for receiving digital asset payments.

Our read

Banks on Jack Henry now have a documented API path for receiving digital asset payments into accounts. The change log does not say which institutions can use it, when, or which networks and assets are supported, so confirm availability for your institution before planning a launch.

Model ML

Oct 22, 2026

ImpactMedium

VerificationVerified

Model ML added a connector to Harmonic's startup discovery data over MCP. Users can search Harmonic for companies, people and investors, enrich target lists, and pull saved searches and their team's network without leaving Model ML. Sourcing screens, market maps and target profiles can be built on live Harmonic data with no exports, and the connector is live for all users with a Harmonic license.

Our read

Coverage bankers, corporate development and private markets teams that already pay for Harmonic can have the agent work from current private company data instead of manual exports. Firms without a separate Harmonic license gain nothing from this release.

Outmarket AI

Oct 22, 2026

ImpactMedium

VerificationVerified

Outmarket AI launched Outmarket Lite, a self serve tier that lets an insurance agency sign up and run its first workflow the same day, with no sales call or implementation project. Lite covers proposal generation, quote comparisons, policy reviews, coverage gap analysis, document verification, ACORD form filling and benefits plan comparisons. It is available now, starting at $1,000 per seat per year.

Our read

Smaller agencies and brokerages can test the core workflows at a published price without an enterprise contract. Automated policy checking, AMS integrations and custom workflows stay in the Enterprise tier, so check which workflows a trial actually covers.

S&P Global Market Intelligence

Oct 22, 2026

ImpactMedium

VerificationVerified

S&P Global Market Intelligence added an MCP server to iLEVEL, its private markets platform for portfolio monitoring, valuations and reporting. Customers can connect iLEVEL data and capabilities to AI tools such as Copilot and Claude while keeping their existing iLEVEL permissions and access controls. The same release names a native iLEVEL Agent and delivers third party valuation results from S&P Global Private Market Valuations directly inside iLEVEL.

Our read

Private credit and private equity teams can use governed portfolio data inside the assistants they already license without building a connector. The release gives no setup or availability detail, so confirm which iLEVEL data the server exposes and how entitlements carry over.

Sigma360

Oct 22, 2026

ImpactHigh

VerificationVerified

Sigma360 launched an Enhanced Due Diligence Agent, available now in its platform. The agent gathers internal and external data, summarizes findings across configurable modules such as source of wealth and funds, beneficial ownership and geographic risk, flags risk factors and mitigants, and drafts a fully cited narrative for regulatory review. Sigma360 says it is designed to cut case handling time by up to 87%.

Our read

The analyst stays the decision maker, and customers can read the agent prompt for each step, which gives model risk and exam teams something concrete to validate. The 87% and 90% figures are design targets, not measured results at a named bank, so test them on your own case mix.

Socure

Oct 22, 2026

ImpactHigh

VerificationVerified

Socure added Actor Classification to Digital Intelligence, which labels each transaction as driven by a person, a script, a bot or an AI agent. It also flags account aggregators, search crawlers and personal agents such as Meta's Muse, using device, request header and on page signals. New response fields include isAutomated, isAgentic, a low, medium or high script risk and the named operator or agent. It is generally available to Digital Intelligence customers on Web SDK 2.9.0 or later.

Our read

Fraud teams can write RiskOS rules that let known aggregators and customer agents through while stepping up unverified scripts. Because classification runs per transaction, a session a customer hands to an agent midway shows up. Older SDK versions return unclassified, so plan the upgrade.

OneSpan

Oct 21, 2026

ImpactMedium

VerificationPartially Verified

OneSpan released Mobile Application Shielding for Android 9.2.0, which now detects Vulkan layer code injection, the Frida injection loader and Linux container emulators such as VMOS. The Code Injection option is now on by default, so apps get the improved detection once they are shielded again, with no configuration change. The release adds Trusted System Keyboards and Trusted System Screen Readers options that automatically trust system installed keyboards and screen readers, and was tested with Android 17.

Our read

Mobile banking teams get coverage against injection and emulator techniques used to tamper with apps. OneSpan calls 9.2.0 a Feature release and advises Stable track users not to upgrade unless they need it, and the new default changes app behavior, so test before rollout.

AdvisoryAI

Oct 21, 2026

ImpactMedium

VerificationVerified

AdvisoryAI added client email reading to Atlas through its Outlook connection, covering both inbox and sent items. Users can name a client and ask what that client last emailed about, or what was sent to them, with no keyword needed. Atlas reads only emails that include the address of a client already held in AdvisoryAI.

Our read

Firms that connected Outlook before October must reconnect and approve a new email permission, or Atlas keeps seeing the calendar only. Compliance teams should sign off on that permission and check that client records hold the addresses clients actually email from.

Alkami

Oct 21, 2026

ImpactMedium

VerificationVerified

Alkami added CarletonCalcs OriginationPlus to MANTL Loan Origination, so loan calculations, contract disclosures, amortization schedules and compliance validation now run inside one lending workflow. The integration supplies disclosure values, fees, taxes and insurance premiums, and checks each loan against federal, state and institutional rules. Where configured, it can automatically adjust loan terms to keep a transaction within defined compliance parameters.

Our read

Lenders on MANTL can drop a separate calculation step and the rekeying between systems. Before turning on automatic term adjustment, confirm how each change is logged and who reviews it, since loan terms can move without a person acting.

Aptus.AI

Oct 21, 2026

ImpactMedium

VerificationPartially Verified

Aptus.AI added a Claude connector to Next-OS that sends relevant legislation and case law from Aptus.AI into Claude, with short verbatim quotes, in force status and a link to the full text. It supports research on a specific provision, judgments by ECLI or citation including the Court of Justice of the EU, and open research from a question. Access is read only, through the user's Aptus.AI account and with revocable consent.

Our read

Compliance teams already working in Claude can pull cited, consolidated regulatory text without switching tools, and Aptus.AI says the connector is included in the Enterprise plan with no credit consumption. Aptus.AI retrieves the sources while Claude draws the conclusions, so test where that line falls.

Bretton AI

Oct 21, 2026

ImpactMedium

VerificationVerified

Bretton changed how its AI drafts quality control scorecards, so the draft now reviews the whole case: the narrative, findings, case inputs, transactions and the agent's investigation record. It also draws on the team's past reviews of the same scorecard. The draft lists specific problems first, each backed by a quote from the case, then answers each scorecard question from them, and past reviews carry over when a scorecard is renamed or versioned.

Our read

Reviewers checking agent work start from quoted evidence instead of an unexplained suggested score. Second line and model risk teams should confirm that reviewers still answer each question themselves, since the draft now leans on the team's own past judgments.

Credora

Oct 21, 2026

ImpactMedium

VerificationVerified

Credora published an indicative A+ rating for JPMD, J.P. Morgan's tokenized deposit issued through Kinexys, the first tokenized bank deposit it has rated. The method anchors probability of default on the issuing bank's balance sheet at 0.023%, then applies ten token modifiers. Five carry charges for missing audits, token to deposit reconciliation, key custody disclosure, size and concentration, giving a final annual probability of default of 0.088%.

Our read

Banks issuing or holding tokenized deposits get an itemized grade on the same scale as the stablecoins and Treasury funds Credora rates, plus the disclosures that would move it. The rating is indicative, scored without the issuer's participation, and Credora states it is not a nationally recognized statistical rating organization.

DataVisor

Oct 21, 2026

ImpactHigh

VerificationPartially Verified

DataVisor launched Vera Code at Money20/20 USA, an AI agent that builds fraud and AML programs on its platform through a conversational interface. It takes a customer's requirements and builds a full solution or single parts of one, including event schemas, aggregated signals, rules, scorecards, alert queues, review workflows and test data, and tests each build before it goes live. DataVisor reports it delivered a real time fraud solution for a Fortune 500 insurer in four days.

Our read

Later changes are verified and approved before they take effect, but fraud and model risk teams should still ask who approves each change and how it is logged. The speed figures are DataVisor's own and come from unnamed customers.

Encompass Corporation

Oct 21, 2026

ImpactHigh

VerificationPartially Verified

Encompass Corporation, the corporate KYC software provider, is now certified to ISO/IEC 42001, the international standard for an AI management system. The standard covers how a company governs the AI it develops, provides or uses. Encompass says the AI management system runs inside the same framework as its existing ISO 27001 information security management system.

Our read

Banks reviewing KYC vendors can ask for audited evidence of AI governance alongside security certificates. The post does not name the certification body, issue date or scope, so request the certificate and scope statement to confirm the KYC platform is covered.

Marloo

Oct 21, 2026

ImpactLow

VerificationVerified

Marloo launched Skills, saved instructions that run a repeated task in Ask Marloo with one command, using the client, meeting and document context Marloo already holds. Users can build a skill by typing /create-skill, by turning a chat into a skill, or from uploaded process documents, and each skill starts private until shared with the team. Marloo says more than 50 firms tested the feature and built nearly 300 skills.

Our read

Firms can turn review preparation, follow up emails and fund comparisons into one shared standard, which also gives compliance a fixed set of instructions to review. Anyone a skill is shared with can run and improve it, so firms should decide who may edit shared skills.

Ravelin

Oct 21, 2026

ImpactMedium

VerificationVerified

Ravelin launched bot activity detection for payment service provider customers, separating unattributed bots from legitimate AI agents at checkout. Transactions with non human traffic now carry Bot detected or Suspected bot labels by confidence level, shown in a Bot details block in the transaction view. PSPs can also write rule conditions on detected bot activity and break down transaction volume by bot status in Insights.

Our read

PSPs can act on scripted checkout traffic across their merchants without building their own bot signal. Detection runs through RavelinJS 2.3.0 or above on the website, so check how many merchants run that version before relying on the new rules.

RavenPack

Oct 21, 2026

ImpactHigh

VerificationVerified

RavenPack released a dataset generation pipeline on Bigdata.com that turns a research hypothesis written in plain language into a labeled dataset ready for backtesting, built from retrieved news and earnings call transcripts. Labeling runs on NVIDIA Nemotron 3 Nano 4B, fine tuned on labels from Nemotron 3 Super 120B, and RavenPack reports it labeled 845,220 excerpts in about 49 minutes on two NVIDIA H200 NVL GPUs. The pipeline is available today to Bigdata.com clients, with a companion notebook.

Our read

Because the labeling model has open weights, RavenPack says regulated institutions can run it in house and inspect its labeling logic, which helps model risk review. The fidelity, cost and backtest figures come from RavenPack's own testing, so ask for the held out evaluation before relying on them.

Renew Risk

Oct 21, 2026

ImpactHigh

VerificationVerified

Renew Risk extended its severe convective storm catastrophe model for solar farms to all 48 contiguous US states, covering hail, tornado and straight line wind. Renew Risk says it had first prioritized hail modeling for the highest risk states. The model uses high resolution hazard data and asset level vulnerability for panel construction, tracker systems and stow logic, and is available through the Global Onshore Renewables Product Suite.

Our read

Insurers, reinsurers and brokers holding US solar risk can now price and check accumulation for all three storm perils in any mainland state. Renew Risk says tornadoes and derechos caused several of the largest severe storm losses at solar farms, so portfolios previously modeled for hail alone are worth rerunning.

Rulebase

Oct 21, 2026

ImpactLow

VerificationVerified

Rulebase now lets teams share Insights reports with named people or teams, choose who else in the organization can open them, and export them as a Word document, PDF or Markdown. Users can select part of a report's instructions and have Instruct, its AI assistant, edit only that passage. When asked for a report, Instruct now drafts it in chat for review before asking whether to schedule it.

Our read

QA and compliance leads can hand recurring findings to reviewers who do not work in Rulebase as standard files. Agents also see reports shared with them on their performance page, so set who can open each report before sharing QA findings widely.

Socotra

Oct 21, 2026

ImpactMedium

VerificationVerified

Socotra added Policy Continuations in beta, which handles a policy change by issuing a new continuation policy built from the existing policy's data. A Continuation Plugin builds the new quote from a transaction with the continuation category, and the new policy can sit on a different product type. Socotra lists cancel and rewrite, transfers between accounts, product upgrades, flat cancellation and changes to fixed data such as timezone as supported uses.

Our read

Carriers can move policies onto a new product version or fix data a normal endorsement cannot touch. It is beta and Socotra asks customers to contact it before production use, and existing Cancellation Plugin code may need changes to avoid charges such as retention fees.

Vouched

Oct 21, 2026

ImpactMedium

VerificationPartially Verified

Vouched updated its Checkpoint SDKs for Express and Next.js in version 1.11.0 so that each accepted signed request from an AI agent is claimed once before it reaches the application's handler. Claims are recorded in the Checkpoint dashboard's shared ledger when an API key and project ID are set, or in a store inside the running process otherwise. A new requestClaimStore option chooses which store is used.

Our read

A captured signed agent request can no longer be resubmitted to repeat the action it authorized. Without an API key and project ID the claim is held by one process only, so sites running several servers should confirm the shared ledger is configured.

Incognia

Oct 20, 2026

ImpactHigh

VerificationVerified

Incognia launched Real Time Scam Prevention for banks, which flags likely scam payments before money moves by checking for signs of a real world relationship between sender and recipient, drawn from apartment level location intelligence and persistent device recognition. It combines that with other risk signals and pools intelligence across participating banks without any of them seeing another's location data. Incognia reports identifying about 70% of scam transactions in an analysis at a bank with more than 80 million active users.

Our read

Authorized push payment scams pass device and login checks because the real customer sends the money, so a sender to recipient relationship signal gives banks a new input for stopping payments. The 70% figure comes from an unnamed bank with no published method, so buyers should ask for validation detail.

Abrigo

Oct 20, 2026

ImpactMedium

VerificationVerified

Abrigo launched Customer Hub inside its financial crime platform, bringing customer, account, transaction, alert, case and risk information into one investigation workspace. It adds a search bar that works from anywhere in the platform and a single view of each customer relationship, so AML and fraud teams work from the same record.

Our read

Investigators at banks and credit unions can connect fraud activity to AML cases without switching modules. It is for Abrigo hosted customers only, and existing hosted clients receive it in waves over the coming months.

Blooma

Oct 20, 2026

ImpactLow

VerificationVerified

Blooma added a GPR Method setting to residential rent rolls and commercial detail rent rolls that sets how Annual GPR is calculated, from actual, borrower or market rents or two hybrids. Cash flows linked to a rent roll now receive Annual GPR and a real economic vacancy deduction, instead of Total Revenue and a $0 vacancy. Existing links moved to the Hybrid Actual & Market method.

Our read

Values on cash flows already linked to rent rolls may change with no user edit, so credit and model validation teams should recheck existing deals. Each formula draws on rows shown in the rent roll table, so reviewers can trace Annual GPR to a visible cell.

Cardo AI

Oct 20, 2026

ImpactMedium

VerificationVerified

Cardo AI made its document extraction engine a data source inside its onboarding platform. PDFs, scanned reports and free form Excel files, such as servicer reports, borrowing base certificates and capital account statements, are turned into typed rows. Those rows then pass through the same data quality checks, transformation logic and audit trail as a loan tape, and each figure links back to the page and table it came from.

Our read

Private credit and asset based finance teams can stop rekeying reporting packs before covenant tests and investor reporting. An analyst verifies each extraction before it is published to the pipeline, so test how corrections carry over to the next file from the same counterparty.

CUBE

Oct 20, 2026

ImpactMedium

VerificationVerified

CUBE published a case study on CMC Markets, which replaced manual horizon scanning with CUBE's Intel regulatory intelligence product across its offices globally. CUBE reports that CMC Markets now spends 50% less time on regulatory monitoring. One global license covers 49 issuing bodies, and the process is described as automated and repeatable in a way the firm can evidence to regulators and auditors.

Our read

This gives compliance teams a named trading firm to reference when sizing the effort saved by automated horizon scanning. The 50% figure is vendor reported with no stated method, and the customer quote speaks to adoption rather than the number, so ask for the baseline.

Fiddler AI

Oct 20, 2026

ImpactMedium

VerificationVerified

Fiddler added versioning to evals datasets in Release 26.24. Every change to a dataset's items now creates a new numbered version, earlier versions stay readable, and a version picker shows what each version created, updated and deleted. Experiments are tied to the dataset version they ran on, so later edits never change their results, and the REST API returns that dataset_version.

Our read

Model validators can rerun an LLM evaluation against the exact test set it first used and show which version produced a result. Item editing includes conflict protection when two people save the same item, so teams sharing a test set do not overwrite each other.

Oscilar

Oct 20, 2026

ImpactHigh

VerificationVerified

TransUnion launched Synthetic Fraud and Credit Washing Scores that carry adverse action reason codes, and named Oscilar's risk platform as the first to incorporate them. Within Oscilar, the scores and their reason codes run alongside credit decisioning in one workflow, so a lender can act on synthetic identity or credit washing risk at account opening and cite it on an adverse action notice alongside the credit reasons.

Our read

Lenders can decline on synthetic identity risk without sending every case through manual review, and can cite credit washing as a reason for the first time. TransUnion describes the scores as being integrated with Oscilar, so confirm they are live in your environment, and have compliance review the reason code mappings and notice language before automated declines go live.

Persona

Oct 20, 2026

ImpactMedium

VerificationVerified

Persona released iOS Inquiry SDK v4.0.0 with PersonaInquirySDK2Protected, a hardened build of the SDK with runtime app protection built in. An app links it in place of the standard PersonaInquirySDK2 product, published from the same package at the same version, with no source changes. The release also adds InquiryEvent reporting for Government ID, Document and Selfie camera and upload actions, camera failures and step transitions.

Our read

Banks and fintechs embedding identity verification in iOS apps can add runtime protection to the capture flow without code changes. The changelog does not say what it detects, so ask Persona before counting it as a security control, and remove PersonaOpenSSL when upgrading.

Sift

Oct 20, 2026

ImpactHigh

VerificationVerified

Sift released Review Agent, which reads a fraud case and gives the analyst color coded risk and trust signals, a short plain language explanation and a suggested action of Block, Accept or Watch. It appears in User Details, generates automatically in Review Queues and can be generated on demand in Explore. Sift says customers report spending 15 to 20% less time on manual review on average.

Our read

Every Sift AI offering starts in assistant mode with the analyst making the final call, and teams can later choose which scenarios to automate. Fraud and model risk teams should decide who approves that move and how suggested decisions are tracked against outcomes.

Vertafore

Oct 20, 2026

ImpactLow

VerificationVerified

Vertafore made its Velocity Coach Agent available in AgencyOne with AMS360 and Sagitta. The agent answers product and how to questions in natural language from Vertafore's product documentation and, when configured, an agency's own standard operating procedures. Users see the supporting sources and can ask follow up questions without leaving the management system, and Vertafore reports that in internal testing it cut the time to find answers by up to 88%.

Our read

Agencies can ground an in workflow assistant on their own procedures, which bears on training time and process consistency. The 88% figure comes from Vertafore's internal testing, so ask for the method and for how uploaded procedure content is stored and kept current.

Incode Technologies

Oct 19, 2026

ImpactHigh

VerificationVerified

Incode made TrustGraph available in production on 19 October as a module that can be added to Flows and Workflows. It checks whether identity data from a session has appeared before, in the customer's own history or across Incode's network, returns signals in real time during the session, and links connected sessions in a Fraud Ring Explorer.

Our read

Banks and lenders get repeat and synthetic identity signals drawn from beyond their own customer base at onboarding and login, without pooling customer records themselves. The signal informs the decision, so each institution's own rules still make the call.

Jump

Oct 19, 2026

ImpactMedium

VerificationVerified

Jump released an FMG Suite integration, built on FMG Connect, that surfaces compliance approved FMG content based on what was discussed in a client meeting, for the advisor to review and use in post meeting follow up. The advisor decides whether each piece fits and whether to send it. The integration is available to all Jump users.

Our read

Advisory firms that already license FMG can put approved content into client follow up without a second tool, and the advisor review step keeps supervision over what reaches clients.

RavenPack

Oct 19, 2026

ImpactMedium

VerificationVerified

RavenPack made an official Bigdata.com connector, built on MCP, available inside Microsoft 365 Copilot Chat and Copilot Cowork. It brings company, market and macro data into the conversation, covering news, filings, transcripts, tearsheets, media sentiment and a customer's own uploaded content, with a source attached to each answer.

Our read

Banks and asset managers standardized on Microsoft 365 can switch on sourced financial data inside the assistant their staff already use, without building a custom agent, and reviewers can trace each figure back to a document.

AlternativeSoft

Oct 19, 2026

ImpactMedium

VerificationPartially Verified

AlternativeSoft released a LASSO factor model that picks the few factors explaining a fund's returns from a published library of 25, up from 16 in its previous desktop model. It covers alternative risk premia, equity styles, credit, rates, commodities, currency and managed futures, and reports adjusted R squared, annualized alpha and tracking error. It is available now in FluidView Desktop, with a cloud launch set for 15 November.

Our read

Allocators testing whether a manager's return is skill or factor exposure get a fully listed factor set and significance statistics they can document for an investment committee or model review.

Finastra

Oct 19, 2026

ImpactMedium

VerificationVerified

Finastra launched StandbyX, a SaaS product that lets US banks digitize and automate standby letter of credit processing without implementing its full Trade Innovation trade finance platform. Finastra pitches it at regional and mid sized banks that still handle standby letters of credit manually or semi manually.

Our read

Regional and mid sized US banks get a lighter cloud route to standby letter of credit automation without the cost of a full trade finance platform. No AI capability is described, so this widens segment coverage rather than AI depth.

WTW

Oct 19, 2026

ImpactMedium

VerificationVerified

WTW added semi autonomous agentic capabilities to the AI Assistant in RiskAgility Financial Modeler, its life modeling and financial reporting platform. Given an unfamiliar product, methodology or migration specification, the assistant can interpret the requirements, plan the work, build or adapt the model, run tests, investigate issues and document the outcome for actuarial review.

Our read

Life insurers facing model migrations or regulatory reporting changes can hand multi step modeling work to the assistant while WTW's governance features keep actuarial judgment and approval with the team. WTW published no time savings figures.

Sumsub

Oct 19, 2026

ImpactMedium

VerificationVerified

Sumsub launched a KYA Module for MCP, a filtering layer between a business's AI agents and the data sources they connect to. It screens each agent request, strips personal and other sensitive data before it reaches the agent, keeps an agent activity log, and can require a live human verification when an agent attempts a flagged action. It works first with Amazon Bedrock AgentCore Gateway.

Our read

Banks and fintechs deploying their own agents get a way to cap what those agents may do alone and to require a verified person for high value actions. Support starts with one gateway, so buyers on other stacks should confirm timing.

GReminders

Oct 18, 2026

ImpactMedium

VerificationVerified

GReminders added a connector to SS&C Black Diamond, set up by an administrator with Black Diamond API credentials. It pulls household and relationship data, account types, AUM, holdings, asset allocation and performance into AI generated briefs before each client meeting. Advisors can also ask questions of that data in plain language through the Ask Anything assistant.

Our read

Wealth firms on Black Diamond can ground meeting briefs and answers in their portfolio accounting system of record instead of manual lookups. It also means client portfolio data now flows into GReminders, which vendor data reviews should cover.

ICE Mortgage Technology

Oct 17, 2026

ImpactHigh

VerificationVerified

The Encompass 26.2 October Service Pack adds support for multiple credit score models across Encompass desktop and web, Consumer Connect, TPO Connect and ICE PPE. Lenders can process credit reports carrying both FICO Classic and VantageScore 4.0, price both at once, and have credit scores and credit disclosure data update to match the model chosen in the rate lock. FICO 10T can be switched on once it is approved.

Our read

Fannie Mae and Freddie Mac have accepted VantageScore 4.0 on conventional loans since 9 September, and Encompass lenders can now pick the score model loan by loan with a record of which one drove pricing and the borrower disclosure. Using it takes configuration and business rule changes, which ICE documents in a separate guide.

Pulse360

Oct 16, 2026

ImpactHigh

VerificationPartially Verified

Pulse360 replaced its homepage with a notice that it has shut down, apologizing to the financial advisors who used it, its investors and its team. Its product, pricing and about pages now return errors, so its AI meeting notes and client deliverables tool for advisors is no longer available. The date is taken from the homepage's last modified record, as the notice itself is undated.

Our read

Advisory firms using Pulse360 need a replacement and should ask what happens to stored client meeting notes, since the notice gives no export, retention or deletion terms.

Facctum

Oct 16, 2026

ImpactMedium

VerificationVerified

Facctum reported that its watchlist management service processed all 1,653 listings in the EU's 22nd Russia sanctions package, adopted on 12 October across three Council Implementing Regulations, in 4 hours and 53 minutes. That time covers validating the changes and making the updated records available for customer screening.

Our read

List update speed decides how long an institution is exposed to newly designated parties before its screening can catch them. The stated time gives buyers a figure to compare across list providers, though the post does not say when the clock started, so ask before writing it into a service level.

WorkFusion

Oct 16, 2026

ImpactMedium

VerificationVerified

WorkFusion's Tara 4.1.1, its payment sanctions screening agent, now suspends a learned decision pattern when an analyst escalates a matching hit, instead of only refusing hits while the escalations continue. The pattern resumes only after a set number of consecutive analyst releases. The release also adds report columns that trace each reapplied decision back to the historical alerts behind it, and detects payment message formats automatically.

Our read

This narrows when the agent may clear payment alerts on its own from past analyst decisions and lets recent human escalations override that automation, a control examiners ask about for autonomous dispositions.

StandardC

Oct 16, 2026

ImpactMedium

VerificationVerified

StandardC Enterprise now supports single sign on through OpenID Connect, with named support for Microsoft Entra ID, Okta, Google Workspace and Auth0. SSO is optional at first; once every user has moved over, an institution can ask StandardC to enforce it and retire StandardC passwords, so an employee disabled in the identity provider can no longer sign in.

Our read

Banks and credit unions can bring StandardC under their existing MFA, conditional access and offboarding controls, with authentication handled entirely in their own identity provider.

BlueFlame AI

Oct 15, 2026

ImpactHigh

VerificationVerified

Blueflame now runs inside Grata Intelligence, the agentic M&A workspace Datasite made available on 15 October, with direct access to Grata's private company data alongside firm knowledge and connected sources. When a target moves into diligence, authorized users can query live documents in the deal's Datasite data room from the same workspace, and each user's existing project permissions limit what the agent can retrieve. A separate connector for outside AI assistants, Datasite Connect, follows on 4 November.

Our read

Deal teams at private equity firms, private credit funds and banks get sourcing data, firm knowledge and permissioned data room access in one agent workspace instead of three separate tools. Permission inheritance from the data room is the control compliance teams will want to test before rollout.

Boost.ai

Oct 15, 2026

ImpactHigh

VerificationVerified

Platform version 14 moves boost.ai's core classification features onto Presis, its own native language model, replacing the mix of GPT-4o and GPT-4.1 models those features used before. Agent orchestration, few shot predictions, out of scope detection, the hallucination guardrail and AI review switch to Presis automatically on update, and customers can no longer choose a different model for them. Boost.ai says the move lowers token cost and latency by taking a third party provider out of the path.

Our read

Banks and insurers running boost.ai assistants now have a vendor built model, not an OpenAI model, making routing and guardrail decisions, and bring your own model is no longer available for those features. Model inventories and third party risk files need updating before the version 14 upgrade.

Built Technologies

Oct 15, 2026

ImpactHigh

VerificationVerified

Built's CLA Draw Agent now lets draw processors and approvers override a blocking finding by choosing a decision and a reason, with an optional comment. The original finding stays visible beside the override rather than turning green, and each override is recorded on the finding with who made it, when and why. Policies the agent judged not to apply to a draw are now listed with its reasoning, so it is clear they were evaluated rather than skipped.

Our read

Construction lenders get an attributable record each time staff proceed against the agent, which is what examiners and model risk reviewers will ask for when an AI can hold up a loan draw.

Hawk

Oct 15, 2026

ImpactMedium

VerificationVerified

Hawk launched a Regulatory Reporting Agent that turns a closed case into a draft suspicious activity report for an analyst to review, mapping fields to each jurisdiction's requirements and drafting the narrative in the institution's own structure. It runs inside Hawk's Case Manager, keeps the four eyes workflow, tracks regulatory deadlines, and checks every field and narrative against the case passage it came from, leaving out anything it cannot verify.

Our read

Hawk's agents now reach the SAR filing step, with an analyst review built into the workflow and a source check on every field, which gives BSA and model risk teams a defined human checkpoint. Hawk's claim that the design eliminates hallucinations is its own, so validators should test the source check on their own cases.

Mortgage Capital Trading

Oct 15, 2026

ImpactMedium

VerificationVerified

MCT released a data integration between its secondary marketing platform and Dark Matter Technologies' Empower loan origination system, available now to mutual clients. The first phase connects Empower loan data to MCT and writes post commitment data, including loan sale commitments executed on MCT Marketplace, back to Empower automatically.

Our read

Lenders on Empower no longer need to rekey or reconcile loan sale data between origination and secondary marketing, a common source of mismatched records. The saving of up to 3 minutes per loan in MCT's headline is its own estimate, not a measured client result.

Oz Forensics

Oct 15, 2026

ImpactHigh

VerificationVerified

Oz Forensics' iOS SDK 10.4.0 deprecated the on device and hybrid liveness analysis modes and set their removal for 1 January 2027, telling customers to move to server based analysis. Android SDK 10.4.0 followed with the same deprecation a day later.

Our read

Institutions that chose on device analysis to keep biometric video on the handset must route captures to an Oz API server, hosted or on premises, by January, which changes data flows, latency and infrastructure plans.

Proof

Oct 15, 2026

ImpactHigh

VerificationVerified

Proof put x401 into production, an open protocol that lets a website ask an AI agent to prove who it is acting for and what that person approved. At its center is an agent identity wallet that holds the person's verified legal identity and lets them approve, on their own device, exactly what the agent can do, such as authorizing a wire transfer or signing a payment mandate. Proof works with Meta's Muse, ChatGPT, Grok, Claude and other assistants that support MCP.

Our read

A bank receiving an agent initiated payment can confirm the customer's verified identity and check the signature on the transaction itself, which gives fraud and dispute teams independent evidence of what the customer approved.

Bits Technology

Oct 15, 2026

ImpactMedium

VerificationVerified

Bits Procedures can now pause while waiting for an outcome, such as a provider check, and resume automatically, with executions showing a Waiting status and timings. Custom screens now run inside procedure executions to guide reviewers step by step, and transaction monitoring rolling windows now go down to the minute. The same release adds Malta as a supported onboarding market.

Our read

Compliance teams can build checks that depend on outside providers and reviewer input inside one tracked workflow, and minute level windows allow tighter velocity rules on fast moving transactions.

Bloomberg

Oct 15, 2026

ImpactHigh

VerificationVerified

Bloomberg made ASKB, the Terminal's conversational AI, generally available to financial professionals. Once enabled, it can also answer questions over a client's own Instant Bloomberg chats, drawing on three months of history, plus the securities in their Worksheets and files uploaded to Bloomberg Drive, with each answer cited back to the source chat or file.

Our read

Firm material such as trader chat, watchlists and internal research now feeds the same AI answers as Bloomberg's data, so banks and asset managers need to review permissioning, retention and supervision before switching it on. Citations on every answer give compliance teams a trail to check.

Dynamic Planner

Oct 15, 2026

ImpactMedium

VerificationVerified

Dynamic Planner launched an integration with Plannr, the UK adviser CRM and back office platform. Client information held in Plannr now flows into Dynamic Planner for profiling, planning and suitability work, and risk profiling results flow back into Plannr. Further integrations are planned.

Our read

Advice firms that run Plannr as their client record can stop rekeying client data into the risk and suitability workflow and keep risk results consistent in both systems. The first phase covers client details and risk results only.

FIS

Oct 15, 2026

ImpactHigh

VerificationVerified

FIS made its AML Investigations Agent, built with support from Anthropic and design partner Amalgamated Bank and running on Anthropic's Claude models, commercially available to any US bank, including banks that are not FIS clients. The agent compares each alert against the customer's profile, history and expected behavior and assembles an evidence summary with its reasoning for the analyst. It runs in FIS's private cloud and integrates natively with the Horizon and Systematics cores, with IBS planned next.

Our read

No case can be closed without explicit analyst confirmation, which gives BSA and model risk teams a defined oversight boundary. Banks outside the FIS base can now buy it, but native integration covers only FIS cores today, so they should ask how it connects to their own core and case tools.

Trulioo

Oct 15, 2026

ImpactMedium

VerificationVerified

Trulioo launched Business Risk Horizon, a business risk scoring capability that combines its KYB data with agentic research into a company's operations and relationships. It traces ownership through layers of corporate entities to flag sanctioned owners or ties to restricted jurisdictions, and every score comes with a plain English explanation and the evidence behind it.

Our read

Payments firms and marketplaces vetting merchants get reason backed scores they can document and defend in due diligence. Trulioo says wider financial services coverage is planned, so banks should confirm availability for their use cases.

Vertafore

Oct 15, 2026

ImpactMedium

VerificationVerified

Vertafore made its Velocity Policy Intake and Policy Check agents generally available, following their early adopter release in September. Policy Intake extracts data from policy documents and flags what is new, changed, missing or conflicting for the user to approve before it updates the agency management system. Policy Check compares the issued policy against that system and its source documents, recommends corrections for approval, and records the review history. Both are live in AMS360; Policy Intake is still in early access for Sagitta.

Our read

Agencies get AI that writes to the system of record only after a person approves each change, with a review trail that bears on errors and omissions exposure. The time savings Vertafore cites come from early adopter testing.

Wink

Oct 15, 2026

ImpactMedium

VerificationVerified

Wink face and palm checkout is now integrated into PAX PxRetailer, the PAX application for enterprise and multilane checkout, and PxDesigner, its screen design tool. It is ready today on every camera enabled PAX device, and merchants can design their own enrollment and checkout screens.

Our read

Acquirers and merchants already running PAX terminals can add biometric checkout on hardware in the field without a separate integration project.

Rulebase

Oct 15, 2026

ImpactMedium

VerificationVerified

Rulebase added routines, automations that follow instructions written like a runbook and run on a schedule, an event, or on demand. A routine can use every Rulebase tool, chosen skills and connected MCP servers, so it can post to Slack or update another system when it finds something. Each run is logged as Acted or Quiet with its version and duration, and edits stay in draft until published.

Our read

Customer operations and compliance teams can have QA and complaint checks run and act without anyone starting them, so buyers should confirm who may publish routines and how the run log is reviewed, since a published routine goes live straight away.

SEON

Oct 15, 2026

ImpactMedium

VerificationPartially Verified

SEON published data handling and security terms for its MCP connector, which lets AI assistants query a SEON account. SEON says it keeps no copy of data passing through the connector and does not train on it, and that the connected AI provider is the customer's own recipient rather than a SEON sub processor. Each connection signs in as a specific user with OAuth 2.1 and inherits that user's permissions, every AI query is written to the Activity Log, and controls to withhold specific data points are not yet available.

Our read

Banks and fintechs assessing the connector as a third party risk now have written terms on retention, training and sub processor status to cite. They also learn that customer data can reach the AI provider they choose, so the transfer terms with that provider fall on the institution.

Shortcut

Oct 15, 2026

ImpactMedium

VerificationVerified

Shortcut 1.4.2 made recurring workflows available, so users can set its Excel agent's tasks to run on a schedule instead of starting them by hand each time. After a scheduled print run, Shortcut now writes reported actuals back into the model automatically.

Our read

Finance teams can keep models current without a person starting each run, which makes it more important to know what the agent changes on its own and how those changes are logged.

S&P Global Market Intelligence

Oct 15, 2026

ImpactMedium

VerificationVerified

S&P Global made Visible Alpha's broker sourced consensus estimates, financial line items and segment data available through its AI Data Portal, which serves S&P data into MCP enabled AI applications. Customers can build AI workflows for earnings preparation, portfolio monitoring and equity research on that data.

Our read

Equity research and portfolio teams at banks and asset managers can pull broker level consensus into their AI agents next to other S&P data, with answers tied to sourced figures rather than model recall.

Bretton AI

Oct 14, 2026

ImpactMedium

VerificationVerified

Bretton changed the related entity check in onboarding screening for cases where a screening provider's match profile merges different organizations with similar names. The check now works out which organization is the customer, and returns no result when it cannot tell rather than reporting a false mismatch. The change applies automatically, with nothing to switch on.

Our read

Teams that feed related entity results into screening rules should see fewer false mismatches, and need a path in rule logic and validation for the new no result outcome.

GBG

Oct 14, 2026

ImpactMedium

VerificationVerified

GBG released the Biometric Onboarding module for GBG Go to production, combining face matching and liveness detection to confirm identity at onboarding. A second variant adds injection attack detection using an encrypted selfie. Journeys can also link the consent screen to the institution's own agreement document instead of GBG's biometric notice, and the agreement's address is stored on the consent record.

Our read

Banks and lenders onboarding remotely get face match, liveness and injection attack defense as one production module, which bears directly on deepfake and virtual camera fraud. Recording the institution's own consent document supports biometric privacy and audit duties.

MyComplianceOffice

Oct 14, 2026

ImpactMedium

VerificationVerified

MCO extended its personal trading and digital asset controls to employee activity on prediction markets, so firms can manage that trading under the same oversight rules. It also widened its AI assisted communications surveillance to more collaboration platforms, naming Telegram and Slack.

Our read

Broker dealers, asset managers and banks now have employees trading event contracts outside normal brokerage feeds, and this brings that activity and chat app messages into the same supervision workflow rather than a separate tool.

Narmi

Oct 14, 2026

ImpactHigh

VerificationVerified

Narmi published its first measured results for AI DecisionAssist, its account opening review agent that recommends a decision for staff to approve or override. Comparing applications reviewed with and without the tool at six institutions over their first six weeks live, median time to decision fell from about 8.5 hours to 4.4, hands on review time fell from 8 to 12 minutes to under 2, and 90% of decisions at all six now finish within three days.

Our read

This is one of few published outcome measures for an agentic decisioning tool in community bank account opening, with a stated comparison method. The results are vendor measured and the institutions are unnamed, and staff still make every final decision.

Oscilar

Oct 14, 2026

ImpactHigh

VerificationVerified

Oscilar made its MCP Server generally available, so Claude, Claude Code or any MCP compatible assistant can connect to an Oscilar environment, turn a written credit or fraud policy into a draft workflow, test it, and work cases. The assistant runs under the user's existing roles, permissions and audit trail, and any workflow change it makes is saved as a draft that someone on the team deploys. Oscilar AI, the in product assistant that explains decisions rule by rule, went generally available at the same time, and both are included at no additional cost.

Our read

Risk and compliance teams get a way for AI assistants to write fraud and credit policy logic without being able to push changes live, the kind of change control a model risk review looks for.

Persona

Oct 14, 2026

ImpactMedium

VerificationVerified

Persona added a Review Matches endpoint that records review decisions on Screening Report matches, either for every matched record at once or for a single record, with a review status and optional risk level and comment. Each review is logged in the report's history, and screening review webhooks now carry a reviewer token that shows whether a person, a Workflow run or an API key made the decision.

Our read

Banks and payments firms that clear sanctions and watchlist hits in their own case tools can write those decisions back to Persona with a record that separates human review from automated review, which supports the audit trail an AML program has to keep.

Aurionpro Solutions

Oct 14, 2026

ImpactMedium

VerificationVerified

Aurionpro released Integro LOS Release 2.0, adding an AI business rule engine that sets lending guardrails from an uploaded credit policy rather than rules built one by one. It also checks documents for tampering and deepfakes, analyzes statements across several bank accounts, maps affordability with an EMI stress view, and routes exceptions to the right authority through a policy based deviation flow with an audit trail.

Our read

Lenders get a shorter path from a credit policy change to live rules, so buyers should ask how generated rules are reviewed and approved before they take effect. Policy routed exceptions with an audit trail give credit risk teams a record to examine.

Axyon AI

Oct 14, 2026

ImpactMedium

VerificationVerified

Axyon AI made an Explainability view live in its Axyon IRIS platform for every client with access to its AI Signals. For each daily signal it uses SHAP based attribution to break the forecast into the drivers pushing a stock up the ranking and those holding it back.

Our read

Asset and wealth managers can see the reasons behind each ranking, which helps them defend positions in investment committees and gives model risk and compliance teams a traceable record of drivers.

Darwinium

Oct 14, 2026

ImpactMedium

VerificationVerified

Darwinium's release 15.3 reweighted the inputs it uses to derive device signatures, its device identification output. Customers should see fewer unrelated devices linked under one signature and fewer cases of one device split across several. On iOS the operating system version had been weighted heavily, which tended to break the link when users upgraded.

Our read

Device signatures feed account takeover, onboarding and payment fraud rules, so banks and fintechs may see linkage counts and rule hit rates shift after this release and should recheck rules and thresholds that depend on them.

Eltropy

Oct 14, 2026

ImpactMedium

VerificationVerified

Eltropy made Branded Texting generally available to credit unions and community banks. Built on RCS, it shows members a carrier verified business profile with the institution's name and logo, and supports rich messages and interactive buttons. It falls back to plain SMS when RCS is unavailable, without the verified profile, and Eltropy's existing consent, opt out and audit controls still apply.

Our read

A verified sender identity on member texts helps institutions counter impersonation and smishing on a channel they already use for alerts and collections, alongside their existing fraud controls. Activation needs sender verification, and the protection is lost on SMS fallback.

Forter

Oct 13, 2026

ImpactMedium

VerificationVerified

Forter launched its Agentic Commerce Engine. Fraud Management now identifies and decides on scraper and AI agent traffic in real time, separating legitimate shopping assistants from fraud and abuse bots, and this part is live for all existing customers. The release adds an AI Activity Dashboard showing which agents transact on a merchant's site, and Agentic Orchestration, which supports headless checkout over agentic protocols through one integration.

Our read

Payment and fraud teams get a decision on each agent initiated order inside their existing Fraud Management flow, instead of blocking bot traffic wholesale. The dashboard shows which agents are buying before a team sets policy for them.

Guidewire

Oct 13, 2026

ImpactLow

VerificationVerified

Guidewire Cloud Platform added cross region support for AWS PrivateLink, so insurers can create inbound and outbound PrivateLink connections when their own AWS VPCs sit in a different region from their Guidewire Cloud VPCs. Guidewire notes changes can take up to three weeks from release to reach each environment.

Our read

Carriers running InsuranceSuite on Guidewire Cloud can keep core system traffic on private AWS networking without moving their own workloads into Guidewire's region, which matters for multi region carriers and network security reviews.

Hebbia

Oct 13, 2026

ImpactMedium

VerificationVerified

Hebbia now integrates with Octus through the Octus MCP Connector, which links every Octus product a firm subscribes to into Hebbia at once. Credit teams can query Octus reporting and legal analysis, private company financials and capital structures, covenant and restructuring data, deal documents and court filings alongside their own deal history, with answers cited to the source.

Our read

Leveraged finance and private credit teams that already pay for Octus can run Hebbia analysis on credit data that public filings do not cover, in one question instead of searching each Octus product. Access follows the firm's existing Octus subscriptions.

IDVerse

Oct 13, 2026

ImpactMedium

VerificationVerified

IDVerse release 1.7.0 lets workspaces set data retention in months and years, with a one year default for new workspaces and a three year maximum for standard retention. Anything longer, including Retain Forever, is available only on request, subject to legal review and IDVerse approval. The same release replaces legacy admin roles with granular permissions across users, records, delivery, flows, product settings and developer tools.

Our read

Retention of ID images, selfies and extracted personal data is a core records question for KYC programs, and institutions whose record keeping duties run past three years now have to request extended retention explicitly.

Kount 360 by Equifax

Oct 13, 2026

ImpactMedium

VerificationVerified

Kount expanded its Email Verification Service model to take in login, new account opening, identity proofing and authorized payment protection events, merging data sets it used to evaluate separately into one global history for each email. As a result First Seen Date moves earlier for many emails, and Last Seen Date, Occurrence Count and Usage Score change too. Kount says existing risk scores and policies are unaffected.

Our read

Banks and lenders that use these email attributes in routing or decision logic outside Kount 360 need to review it, as Kount itself advises. It is also a documented change to model inputs that a model risk team would log.

Primer

Oct 13, 2026

ImpactMedium

VerificationVerified

Primer published a retrieval evaluation of company root, the company based file layout it ships with its research agent, tested against three other layouts and two tool setups over 480 attempts by GPT-6 Sol. With enhanced search, company root surfaced all the required evidence in 60 of 60 attempts at about 9.3 US cents and 41 seconds each, and complete retrieval across all layouts rose from 85.4% with base tools to 97.1% with enhanced tools.

Our read

Research teams get a disclosed method plus cost and time figures for how the agent finds evidence in a firm's own archive. Primer states the limits plainly: the result shows the evidence reached the agent, not that its sixty answers were correct.

Ravelin

Oct 13, 2026

ImpactLow

VerificationVerified

Ravelin launched a documentation MCP server that connects AI coding assistants such as Codex, Claude and Cursor to its developer docs. Besides searching integration documentation, it can validate API request payloads and help troubleshoot integration errors. It is read only and needs no API key or authentication.

Our read

Engineering teams integrating or updating Ravelin's APIs can check payloads and integration guidance inside their coding tools instead of switching to the browser, which shortens build cycles. It touches developer tooling only, not the fraud models.

Aloan

Oct 13, 2026

ImpactMedium

VerificationPartially Verified

Aloan published a public reference for its Commercial API v1, with a downloadable OpenAPI 3.1.0 document and 30 operations covering relationships, deals, term sheet intake and the credit memo. A separate sandbox has its own keys and data, and the API supports IP allowlists.

Our read

A bank's IT team can scope a loan origination, CRM or data warehouse integration and test it in a sandbox before contracting, rather than relying on vendor led engineering alone.

Bdeo

Oct 13, 2026

ImpactMedium

VerificationVerified

Bdeo published first production results from its visual AI damage assessment at UK motor claims handler S&G Response, delivered through the Claim Technology platform. After deployment, 69% of claims were processed within 24 hours of notification through digital self service, 98% were routed automatically to the correct repair pathway, and smart repair identification rose 15%. Early indications put potential savings at £250 to £375 per claim.

Our read

Motor insurers get named production figures on automated routing and cycle time from a live UK deployment to benchmark against. The figures are early and self reported, and the savings are an estimate rather than a measured result.

eGain

Oct 13, 2026

ImpactMedium

VerificationVerified

eGain now integrates with Jack Henry core platforms for banks and credit unions as a member of the Jack Henry Fintech Integration Network, so governed knowledge can move between eGain and the core. The same release grounds AI agent answers in more than 100 regulators and industry sources across North America and EMEA through its Trusted Knowledge Network, and adds prebuilt agentic workflows such as financial assistance screening and eligibility checks.

Our read

Banks and credit unions on Jack Henry cores can connect eGain's AI agents to core data instead of answering questions in isolation, and answers drawn from named regulators give compliance teams a traceable source.

Fiddler AI

Oct 13, 2026

ImpactMedium

VerificationVerified

Fiddler release 26.23 adds support for Portkey AI Gateway. Fiddler guardrails can now redact PII and secrets inline in Portkey traffic, replacing detected items in the request or response so the sanitized call still proceeds, and blocking the request if redaction fails. Fiddler can also ingest Portkey's OpenTelemetry traces without application code changes, using Portkey's experimental log export.

Our read

Banks and insurers that route model traffic through Portkey can mask customer identifiers before prompts reach a third party model and get monitoring coverage without touching each application, which supports data minimization reviews.

FinTech Studios

Oct 13, 2026

ImpactMedium

VerificationVerified

FinTech Studios released version 1.1.0 of its MCP server for the FTS Studio API, adding four tools that write to a user's Studio account: creating collections, updating collection filters, creating insight channels and updating the daily brief. The connector launched in September with read only tools, and the package notes say the AI client asks the user to confirm each write before it runs.

Our read

An AI assistant using this connector can now change saved collections, scheduled channels and briefs rather than only read them, so firms should decide which writes they allow and confirm which version their deployment runs.

TRM Labs

Oct 13, 2026

ImpactHigh

VerificationVerified

TRM Labs launched Ardence, an AI investigation platform for government investigators, cyber threat analysts and compliance teams, available today. It applies TRM's tested investigative methods to pursue more lines of inquiry than a team can by hand, and its data network combines TRM intelligence with more than 60 partner providers of corporate, cyber, identity, sanctions and financial data for cases on fraud, cybercrime and terrorism financing.

Our read

Financial crime teams at banks and crypto firms get an AI layer that takes TRM beyond on chain tracing into wider investigation work in one case record. TRM says investigators direct the work and every finding stays linked to its source, which bears on how the output holds up in review.

Veriff

Oct 13, 2026

ImpactHigh

VerificationVerified

Veriff published Presence Verification, which uses a passive selfie video and device signals from its native SDKs to confirm a real, live, single person is completing the session, without asking for an identity document. A higher tier, Presence Verification with Coercion Detection, also analyzes session audio and video for signs that a third party is guiding the user.

Our read

Third party coaching is a common pattern in money mule and authorized push payment fraud, so a dedicated coercion signal gives bank fraud teams something to route to review or a step up check. Coercion detection needs microphone access and is limited to adult users.

SBS

Oct 13, 2026

ImpactHigh

VerificationVerified

SBS launched SBS AI Foundation, a platform for generative and agentic AI in core banking that banks can run on a private cloud, a managed cloud or fully on premises, so data never leaves their environment. It runs on NVIDIA's stack, with NIM microservices for the AI layer, NeMo Guardrails to keep responses within each bank's regulatory boundaries, and open Nemotron models, and draws on data from SBS core banking, lending, digital banking and regulatory reporting. It is live at Scottish Building Society and now available to SBS's client base.

Our read

Banks under strict data residency rules can run generative AI on core banking data without sending it to an outside AI provider, and the named model and guardrail components give model risk and procurement teams a concrete supply chain to review.

Shufti

Oct 13, 2026

ImpactMedium

VerificationVerified

Shufti added an IBAN Check to its Bank Account Verification API and expanded documented coverage from 48 to 109 countries. The IBAN Check validates an IBAN's structure and check digits and returns the bank behind it, without claiming who holds the account, and responses now say whether a result confirms ownership or only the IBAN.

Our read

Payments, lending and payout teams can pre validate destination accounts in far more countries through one call, but must read the verification mode, because an IBAN only result does not confirm ownership. Existing integrators should retest, since keys the selected check does not take are now rejected.

Spektr

Oct 13, 2026

ImpactMedium

VerificationVerified

When a spektr transaction monitoring rule fires, the case now stores the measured value behind each condition, shown with its threshold and a Met or Not met badge. The AI case summary quotes the same stored numbers, so it cannot cite a figure the rule did not fire on. Rule conditions also now accept arithmetic and named formulas.

Our read

Reviewers, auditors and model validators can see exactly why an alert fired from the case record, and the AI narrative is tied to recorded rule inputs, which lowers the risk of an unsupported figure reaching a SAR.

SS&C Technologies

Oct 13, 2026

ImpactMedium

VerificationVerified

In SS&C Blue Prism WorkHQ, requests from the SS&C AI Gateway to its UK production deployment on AWS now go through a UK based load balancer instead of a US based one. The same release ships a fixed 90 day deletion of files uploaded through WorkHQ form attachments.

Our read

UK banks, insurers and asset managers using WorkHQ's agentic workflows can keep AI Gateway traffic on UK infrastructure, which matters for data residency commitments and outsourcing reviews that previously had to account for a US hop.

Sympera AI

Oct 13, 2026

ImpactMedium

VerificationVerified

Sympera published a case study in which Republic Bank of Chicago reports a 30% increase in new business relationships since 2025 after deploying Sympera across its 20 branches. Market managers start each week with a list of local businesses showing activity and the reason each is worth a call.

Our read

It ties a named bank to growth in new relationships rather than lead volume, giving buyers a reference to check. The figure is the bank's own, with no stated baseline or method.

Temenos

Oct 13, 2026

ImpactMedium

VerificationPartially Verified

Temenos published its 2026 Performance Benchmark report, which reports 16,385 transactions per second in a composable core banking benchmark environment run on Temenos SaaS on Microsoft Azure.

Our read

Banks get a stated benchmark to compare with their own volumes when sizing Temenos SaaS. The report is the vendor's own test, so buyers should ask how its transaction mix compares with theirs.

Carta

Oct 12, 2026

ImpactMedium

VerificationVerified

Carta received authorization from the CSSF, Luxembourg's financial regulator, as a Specialised PSF, adding licensed fund administration and transfer agency in Luxembourg to its platform. Carta can now administer Luxembourg domiciled funds, including SCSp, RAIF, SIF and SICAR structures, on the same system it runs for clients in other jurisdictions.

Our read

US and Asia Pacific managers running parallel or feeder vehicles into Europe can keep Luxembourg administration on one platform instead of reconciling with a local provider. The authorization is announced by Carta, so diligence teams should confirm the entity on the CSSF register.

Caudena

Oct 12, 2026

ImpactMedium

VerificationVerified

Caudena published a benchmark of 11 model configurations running agent workflows on its Prism investigation tools, scored on graph recall, graph precision and required report findings across Bitcoin, Ethereum and TRON tracing cases, with a 25 minute and $3 cap on each attempt. Across 69 scored attempts run 25 to 28 September, Gemini 3.8 Flash at medium reasoning effort scored 100 at an average of 3 minutes 40 seconds and about $0.32 per investigation, tied on quality with Claude Opus 5.5 and GLM-5.3.

Our read

Investigation and compliance teams get a stated method, time and cost figures, and the named models behind Caudena's agents, which feeds model validation. Caudena notes there is no comparison with human investigators and that a perfect score does not prove every narrative claim was supported.

HyperVerge

Oct 12, 2026

ImpactMedium

VerificationVerified

HyperVerge's HyperKYC Flutter SDK 3.5.0 bundles native Android HyperKYC 3.5.0 and iOS HyperKYC 1.11.0, both adding injection checks to the document capture module. The Android release also adds a pre probe step for selfie capture and fixes false rejections in the existing injection checks.

Our read

Lenders and banks onboarding remotely can now screen ID document captures, not only selfies, for injected image fraud in the same SDK flow. Fewer false rejections on Android also means fewer good applicants dropped at onboarding.

ibl.ai

Oct 12, 2026

ImpactMedium

VerificationVerified

ibl.ai gave every agent a privacy policy setting, shipped in v2.12.0 on 12 October: an admin can let users decide, make every conversation with that agent incognito, or never allow incognito, alongside the existing PII filter. One API endpoint resolves the agent policy, the organization setting, the chat toggle and the user default, and returns the resulting mode, which tier decided it and whether it is locked.

Our read

A bank can force an applicant intake agent to keep no chat history or memory while an internal research agent keeps it, instead of one switch for the whole organization. That per agent retention record is what a GLBA safeguards or data minimization review asks to see.

iDenfy

Oct 12, 2026

ImpactMedium

VerificationVerified

iDenfy's Sprint 137 lets bank card verification run as a step of identity verification inside apps built with its iOS and Android SDKs. Capture runs in an iDenfy hosted secure view inside the SDK that sends card images straight to iDenfy's PCI DSS certified card environment, so they never pass through the client app's code. The same release adds the Brazil CPF check to the native SDKs.

Our read

Fintechs and lenders that onboard in native mobile apps can now match a payment card to the verified identity in that flow, without their own app code ever handling card images.

Kore.ai

Oct 12, 2026

ImpactMedium

VerificationVerified

Kore.ai Agent Platform 1.8.0 gives Autoloop, its evaluation driven loop that proposes improvements to agents, three approval modes. Autopilot applies changes automatically, Copilot applies them only after a person approves, and Advisor only suggests them. The mode can be set when a loop starts or changed mid run, and it carries over when a loop restarts.

Our read

A bank can hold agent optimization to Copilot or Advisor so no AI proposed change reaches a live agent without a named reviewer, which fits model change control. Autopilot is now also on the menu, so governance teams should decide which modes they allow.

Middesk

Oct 12, 2026

ImpactMedium

VerificationVerified

Middesk removed the identifier_assessments array from Business Risk assessment responses. An assessment now carries only the overall verdict, score, level and business level risk dimensions, and email risk becomes its own email_risk order. Integrations that read the per identifier scores need to drop that dependency.

Our read

Banks and lenders that parse or document Middesk risk outputs need to update integrations and any model documentation that relied on per identifier scores, and order email risk separately if they used it.

nCino

Oct 12, 2026

ImpactMedium

VerificationVerified

nCino introduced a Loan Officer Productivity Suite inside its Mortgage Point of Sale, bundling six AI tools in one loan officer workspace. At its center is Mortgage Advisor for Loan Officers, which brings nCino's borrower facing conversational AI to originators so they can ask about their pipeline, loan files and GSE guidelines, and have it create tasks or send reminders once the loan officer approves.

Our read

Mortgage lenders get a guideline aware assistant that acts only after loan officer approval, a control worth testing in due diligence. nCino names early adopters already using it but does not state general availability.

Ocrolus

Oct 12, 2026

ImpactHigh

VerificationVerified

Every Ocrolus mortgage customer now gets Ocrolus Detect, its document fraud product, at no added cost. Detect checks bank statements, pay stubs and tax forms for tampering and, through Ocrolus' integration with Resistant AI, for files created or edited with generative AI. Each document gets an authenticity score with reason codes, such as edited text or balances that do not reconcile, and underwriters see the results in the Ocrolus platform or in the Encompass embedded viewer.

Our read

Mortgage lenders already on Ocrolus get document fraud screening without a separate line item, which changes the cost comparison against standalone fraud tools. The reason codes show the underwriter why a document was flagged, and the underwriter still makes the call.

DigiFi

Oct 12, 2026

ImpactMedium

VerificationVerified

DigiFi shipped an AppOne integration, so dealer applications submitted through AppOne are created in DigiFi automatically and lenders can send decisions, funding updates and comments back. A new Aggregators section holds AppOne and Dealertrack in one view for managing indirect lending integrations.

Our read

Lenders in RV, marine, powersports and independent auto can take indirect applications from the AppOne dealer network without a custom build, alongside their existing Dealertrack flow.

Unit21

Oct 12, 2026

ImpactMedium

VerificationVerified

Unit21 released a Corelation KeyStone core integration for credit unions. After a one time historical load, member, account, card and transaction data flows into Unit21 about every five minutes, and analysts can act back in KeyStone through the KeyBridge API from inside an alert or case, for example blocking a card.

Our read

Credit unions on KeyStone can stop fraud in the core without opening a second system, and keep the action in the same case record that feeds their BSA work.

Wealth Dynamix

Oct 12, 2026

ImpactMedium

VerificationPartially Verified

Wealth Dynamix published a fair use policy for its CLMi platform that sets API limits, per user or service account, of up to 6,000 calls over a sliding five minute window and 50 concurrent calls. It caps ingested data storage at 1 TB and sets a fair use ratio of 35 monthly active external portal users per authorized user, and customers can request API usage telemetry once a year.

Our read

Wealth firms building their own integrations on CLMi can size against published throttling, storage and portal limits before signing, rather than discovering them after go live.

Charm Security

Oct 11, 2026

ImpactLow

VerificationVerified

Charm published a privacy policy for its fraud prevention browser extension, distributed through the Chrome Web Store, setting out how the extension handles bank data. It runs only on pages in a permitted list that the customer sets in the Charm admin portal and Charm enforces, and on those pages it reads the case, customer, account and transaction details on screen. IP addresses are trimmed and hashed daily, Charm says it takes steps to filter identifiers out of AI tool inputs while not guaranteeing their removal, and the page data is not used for advertising, profiling or anything unrelated to the service.

Our read

A browser extension that reads case and account screens inside bank tools needs a written scope boundary before GLBA vendor review signs off, and this gives privacy and security teams one to test the deployment against. The AI input filtering is a best efforts commitment, so reviewers should assume customer data can reach the AI tools.

Intapp

Oct 10, 2026

ImpactMedium

VerificationVerified

Intapp published version 1.0.16, the first release of the DealCloud Python SDK 1.x line, which its developer changelog calls a complete rewrite covering schema, records, history, files, users, publications, backups, relationship intelligence and merge APIs. It adds retry handling per error code with backoff, respect for rate limit headers, streaming pagination for bulk reads, and output to pandas or Polars.

Our read

Investment banks and private capital firms that move DealCloud data into warehouses, models or internal tools get a supported client, instead of maintaining their own API wrappers. The 1.x line is a breaking change, so teams on older versions need the migration guide before upgrading.

Middesk

Oct 9, 2026

ImpactHigh

VerificationVerified

Middesk updated its verification agents: the Address Verification and Address Type Risk agents now pull Google Street View imagery from several angles, read exterior signage against the applicant's business name and classify the property as a storefront, office, industrial site, residence, mail drop or vacant land, with the images and assessment embedded in the review panel and audit PDFs. A new Document Verification agent, in early access for select customers, screens uploads such as IRS letters, formation filings, leases and bank statements for tampering and reconciles them against verified records. Teams can now set which agents run automatically and the confidence thresholds that force human review, P95 turnaround fell from about two minutes to one, and a unified audit trail links each finding, decision and action.

Our read

Business onboarding teams can automate the map checks and document reviews analysts do by hand while setting their own thresholds for when a person must look. Middesk's claim that the imagery saves 5 to 10 minutes per review has no stated method.

Quartr

Oct 9, 2026

ImpactMedium

VerificationVerified

Quartr's first party investor relations data is now available inside Model ML through a native MCP integration, live for Model ML users who also hold a Quartr Pro subscription. Users can query live and recorded earnings calls, transcripts with speaker identification, slide decks and filings covering more than 16,000 public companies in 65+ markets, and use them for source traceable memos and decks and for qualitative work such as tracking tone shifts and guidance credibility.

Our read

Analysts already working in Model ML can ground qualitative research in primary company material without switching tools. It requires a separate Quartr Pro subscription.

IDnow

Oct 9, 2026

ImpactMedium

VerificationVerified

Sweden's Trusted List, maintained by the supervisory authority PTS, lists two new qualified eIDAS trust services for IDnow Trust Services AB as granted from 9 October: the IDnow Qualified Electronic Archiving Service and IDnow Long Term Preservation of Signatures and Seals. They join IDnow's existing qualified signature, seal, timestamp and attestation services on the same list.

Our read

Banks and insurers that must keep KYC records for years can archive onboarding evidence with an EU wide legal presumption of integrity and origin, held by the same supervised provider that ran the identity check.

Darwinium

Oct 8, 2026

ImpactMedium

VerificationPartially Verified

Darwinium launched two capabilities aimed at fraud by AI agents. Journey Transition Probability scores each step toward a login, account change or payment against the normal order and timing for that customer, that business and that journey type, flagging steps that look normal alone but suspicious in sequence, whether a person, bot or agent is acting. MCP Protection adds agent tool calls made over the Model Context Protocol to the same journey record as web and mobile activity, verifies agent credentials and can hold payments or other higher risk actions for extra checks, with decisions enforced through the business's existing CDN.

Our read

Fraud teams preparing for agentic commerce can watch what an agent does across a whole journey rather than judging single events. Darwinium's documentation already showed transition probability features in 2025, so MCP Protection is the newer part, and GA status and pricing are not stated.

Backbase Conversational Banking

Oct 8, 2026

ImpactHigh

VerificationVerified

Backbase made Conversational Banking generally available. Customers and staff state a need by voice or text, an AI agent maps it to a banking action such as blocking and reissuing a lost card or checking a payment, and it acts only within bank defined permissions, policies and approvals, with explicit customer confirmation for consequential requests and a record of context, decisions and outcomes. When a case needs human judgment, staff receive the relevant information and a summary of what the agent already did. Backbase says it is in production at more than 10 banks, including a large South African bank that has handled 22 million customer interactions with self service containment rising from 20% to 70%.

Our read

Banks can move from assistants that answer questions to agents that complete routine service tasks inside limits they set, with an audit trail. The containment figures come from an unnamed bank with no stated method; Meriwest and VeraBank, longtime Kasisto customers, are quoted without numbers.

Murex

Oct 8, 2026

ImpactMedium

VerificationVerified

Murex's MX.3 platform can now connect natively to ICE's real time ICE Consolidated Feed through a new IDS-CF-MD interface that streams market data directly into MX.3, available from version 3.1.68. Full streaming requires an entitlement system, either the client's own through Murex's Entitlement SPI or MX.3's native Real-Time Data Access Control System. It builds on ICE integrations in place since 2022 for curves, volatilities and credit and bond data.

Our read

Trading and risk teams on MX.3 can take ICE real time pricing without building and maintaining a custom feed handler. Licensing and pricing terms are not stated.

NICE Actimize

Oct 8, 2026

ImpactHigh

VerificationVerified

NICE Actimize says its solutions have been independently certified to ISO/IEC 42001:2023, the international standard for AI management systems, with the certificate issued by SII-QCD. The stated scope covers the development, provisioning and deployment of its financial crime, risk and compliance software, and the post describes the practices behind it, including explainable and attributable agent decisions, defined agent boundaries with human oversight, full audit trails, and continuous monitoring with change controls.

Our read

Banks running model risk and third party AI reviews can cite an external certification of how NICE Actimize governs its AI rather than relying on questionnaires alone. The post gives no certificate number, issue date or validity period, so ask for the certificate during procurement.

Smart Capital Center

Oct 8, 2026

ImpactHigh

VerificationVerified

Smart Capital Center launched Fraud Alerts, which checks borrower submitted commercial real estate financial and operating data for inconsistencies and possible misrepresentation from origination through asset management. It normalizes Excel workbooks, scanned PDFs and unstructured operating statements, compares documents from the same period (such as a rent roll against an operating statement), tracks the same data points over time to flag unsupported changes in leases, rents or occupancy, and reconciles construction draw requests against budgets and inspection reports. Alerts are tagged by fraud type, linked to source documents with a variance report, and can be prioritized, dismissed or annotated on the asset record.

Our read

CRE lenders can look for phantom occupancy, inflated income and similar manipulation across a whole portfolio rather than in sampled reviews, with a documented trail for each alert. The fraud scenarios on the page are illustrations rather than customer results, and no lender is named as using the feature.

Unit21

Oct 8, 2026

ImpactMedium

VerificationVerified

Unit21 joined Alkami's partner program with a live real time integration: when a member logs in, updates a profile or moves money on the Alkami Digital Banking Platform, Alkami sends the event to Unit21 over a synchronous API, and Unit21 checks it against the institution's rules and returns pass, step up authentication or block before the action completes. Alkami activity lands in the same Unit21 workspace as core and payment data, where alerts flow into case management with AI triage and prefilled SAR or CTR filing reviewed by a person.

Our read

Banks and credit unions on Alkami can stop account takeover and payment fraud at the moment of the digital action and investigate it alongside core data. Unit21 cites results from Kinecta Federal Credit Union, but those describe Unit21 running alongside Alkami before this integration, not the integration itself.

Candescent

Oct 8, 2026

ImpactMedium

VerificationVerified

Candescent's developer changelog documents an MCP server that ships in beta inside its Forge CLI, so coding agents such as Cursor, Claude Code, VS Code, Codex and Gemini CLI can sign a developer in, call Digital Insight APIs, and build, preview and submit widgets and Aspects through chat. The server shares the developer's existing CLI session, and destructive or write operations ask for confirmation. The same release adds browser sign in to forge login.

Our read

Banks and credit unions building extensions on Candescent's digital banking platform can drive API calls and widget submissions from the coding assistants their developers already use. The agent runs on the developer's own Forge sign in, so access control follows the institution's existing developer accounts.

Consilient

Oct 8, 2026

ImpactHigh

VerificationVerified

Consilient launched Federated Synthetic Data, a library of synthetic AML training data covering 24 validated typologies across five risk families and about 110 million labeled transactions. Institutions use it to train detection models for typologies they have few confirmed cases of, such as trade based laundering, corruption and sanctions evasion, then train locally and share only approved model learning through federation. In Consilient's test, models trained on synthetic data alone ranked a real bank's confirmed suspicious accounts 10 to 15 times better than random.

Our read

Banks short on confirmed cases for a typology can run model based detection from day one instead of waiting years for labels, without moving customer data. The test bank is unnamed and the result comes from Consilient's own study, so validators should ask how the bank's customer mix compares with theirs and request the method behind the figure.

Keylane

Oct 8, 2026

ImpactHigh

VerificationVerified

Keylane launched Keylane Automate, a platform where insurers and pension funds build their own AI agents and workflows for claims, underwriting, member communications and case handling in a no code environment. Customers choose the AI models they deploy and can use prebuilt agents or build their own with a developer SDK. Keylane says higher risk decisions go through human review, every decision is auditable by the customer, and data is hosted on European infrastructure.

Our read

Insurers and pension funds running Keylane's core systems can put AI agents directly on governed policy, claims and member data instead of building separate integrations. The audit trail and human review bear on the EU AI Act's high risk rules for life and health insurance pricing and risk assessment, which the Digital Omnibus deferred to 2 December 2027, so buyers have time to test those controls before they become binding.

Alkami

Oct 8, 2026

ImpactMedium

VerificationVerified

Alkami's Fall 2026 platform release lets authorized institution staff manage PGP encryption themselves for data extracts delivered from the Alkami data lake. Business accounts now need an authorized approver before a new authorized user goes live or an existing one is removed, and staff with read only access can view business wires and admin without changing them. The release also adds SWIFT and IBAN validation before international payments, and Mitek and Unit21 join as Alliance partners.

Our read

Institutions can control encryption on sensitive data leaving the platform themselves rather than through a vendor request. The approver step closes a common gap in business account administration.

Cardo AI

Oct 7, 2026

ImpactMedium

VerificationVerified

Cardo AI added PDF report generation that turns a deal dashboard into a paginated, presentation ready report tied to live deal data. Users set the layout, commentary, branding and cover page, choose which tables, KPIs and charts to include, and save several templates per deal for different audiences, such as a monthly investor report and a covenant pack; reports can be generated for every deal in one batch or scheduled to email recipients on a reporting date, and each report is logged with its producer and template version.

Our read

Asset based finance teams can cut the manual formatting behind investor reports and covenant packs and standardize them across deals. Exports follow existing user permissions.

Kore.ai

Oct 7, 2026

ImpactMedium

VerificationVerified

Version 1.7.3 of Kore.ai's Agent Platform changes how agents handle several requests in one message: two new settings decide when a request counts as finished and have the supervisor capture every request at once, and the automatic strategy now handles requests one after another and asks the customer when they are ambiguous. When a tool call needs approval, customers now see a plain question built from readable labels such as account and amount, with sensitive values masked and internal instructions hidden. Tool hooks can pass the model a short summary instead of a large result, and project imports keep their model selections.

Our read

Banks running customer facing agents get clearer consent prompts before an agent moves money or changes an account, and fewer dropped or misordered requests. Teams that relied on parallel handling of requests now have to set it explicitly.

MSCI

Oct 7, 2026

ImpactHigh

VerificationVerified

MSCI launched SignalLab, a platform of more than 600 research based, methodology governed investment signals updated daily, with tools to test and deploy them, spanning Risk Premia, Micro-Industries and MSCI's existing FactorLab and Crowding signals. An AI agent turns research into tested, scored Risk Premia signals that an MSCI researcher validates, and for Micro-Industries a proprietary language model clusters companies while LLMs refine definitions, checked by MSCI's model validation team. Clients can bring their own data into the same environment and test ideas through history.

Our read

Quant and fundamental teams get vetted signals and a place to test them against their own data, with human validation of what the AI produces. Fundamental Momentum and AI Exposure modules are planned, and MSCI gives no access tiers or backtest results.

Shortcut

Oct 7, 2026

ImpactMedium

VerificationPartially Verified

Shortcut release 1.2.1 introduced Sho, an autonomous agent with its own cloud computer that takes work by voice or text, carries it out in the background and remembers conversations and notes between sessions, in early alpha for organizations with voice mode enabled. The same release moves users to Pivot 0.7, which replaces Pivot 0.6 and which Shortcut describes as much stronger at financial modeling and especially at auditing financial models; existing users are upgraded automatically.

Our read

Analysts can hand longer jobs to an agent that keeps working while they do something else, and teams auditing financial models get a stronger default. Sho is gated twice, by the alpha and by voice mode, which only began rolling out on 25 September, and Shortcut gives no benchmarks for Pivot 0.7.

Socotra

Oct 7, 2026

ImpactMedium

VerificationPartially Verified

Socotra's 7 October release removes the deprecated extended query parameter from all list endpoints, 57 endpoints across eight services led by Billing and Policies, leaving ListPageResponse as the only response shape. It also adds Labels for categorizing tasks in the Operations Workbench, configured per tenant (up to 100, and fixed once deployed) and applied from the Task Dashboard or through the Add Task and Update Task APIs, and it retires the Past Due status, moving those tasks to Active.

Our read

Insurers and MGAs integrated with Socotra need to check any calls that still pass the extended parameter, since this is a breaking API change. Operations teams get a simple way to mark blocked, on hold or pending tasks and filter by them.

Spade

Oct 7, 2026

ImpactMedium

VerificationPartially Verified

Spade's merchant enrichment is now built into Increase's card issuing platform, available to card programs on Increase today. Each card transaction carries a Spade record with merchant name, legal name, logo, website, phone, store location and an industry category up to four levels deep, matched in real time against Spade's Merchant Graph, which links locations to brand and owner, identifies the merchant behind processors such as Square and flags delivery apps and marketplaces. The data is available during the authorization decision and is optional by design, so an authorization never depends on it.

Our read

Card programs on Increase can show recognizable charges, settle questions about who charged a cardholder before they become disputes, and write location and category rules into authorization without a separate integration. Spade's figures of under 40 ms and more than 99% accuracy are its own, with no method given.

AlphaSense

Oct 6, 2026

ImpactHigh

VerificationVerified

AlphaSense released the next generation of SuperAnalyst, the agent it first introduced in early access in June. Users hand it an objective rather than a prompt, and it plans the work, finds and reconciles evidence across filings, broker research, earnings calls, 300,000 expert interviews and a firm's own permissioned content, runs calculations and writes and runs code, producing memos, Excel models, comparison tables, dashboards and board decks. A Context Graph ties each data point to its source, and prebuilt Skills cover tasks such as earnings previews, buyer universe development and competitive mapping, with custom Skills available.

Our read

Analysts can delegate multi step research and modeling with sources attached, which suits investment memos and diligence that must be checked. AlphaSense's help center still lists SuperAnalyst as beta, and scheduled and event triggered monitoring is described as coming soon.

Feathery

Oct 6, 2026

ImpactMedium

VerificationPartially Verified

Feathery added a supervised mode for Computer Agents that pauses before consequential actions, such as submitting an account change or updating a client record in another system, and whenever the agent has a question, resuming once a team member approves, rejects or answers. Agents can replay a recorded run, a new API lets customer systems start a run and read its result, and audit logs now record who viewed or changed forms, submissions, envelopes, settings and field data, including refused requests; Enterprise teams retrieve the log through the API by date range.

Our read

Financial services teams automating work in other systems can keep a person on every consequential step and produce an activity record for compliance reviews. Feathery does not say which plans include supervised mode.

Incognia

Oct 6, 2026

ImpactHigh

VerificationVerified

Incognia launched AI Agent Detection for financial institutions, which identifies agentic activity and assesses the risk of each action an agent initiates separately from the agent itself, combining signed request verification, agent origin classification and bot detection with Incognia's device, location and account signals. When an action needs more confidence, it can be checked against the customer's trusted devices and account history. The same release adds a read only MCP server that lets fraud analysts query evidence and device and account links from compatible AI apps.

Our read

As customers start sending AI assistants to act on their accounts, fraud teams need to judge the action as well as the agent, and this lets them do it with signals they already trust. Incognia does not state GA status or customers; the web behavioral biometrics in the same release were already logged on 23 September.

Majesco

Oct 6, 2026

ImpactHigh

VerificationVerified

Majesco's Fall '26 release, more than 700 capabilities across P&C, life and health, and retirement, introduces Immersive AI, which uses on screen context inside core workflows to surface agent processed work, recommend next actions and complete routine tasks within guardrails, on a framework of Agent Studio, an agent and tool registry, a model catalog and MCP with audit and human approval steps. P&C gains five machine learning models (risk signals, predictive scores, quote acceptance likelihood, document extraction and operational reporting), OFAC screening of claim payment recipients and AI assisted workbook creation from approved filings, while life and health gains six named models and a new Intelligent Underwriting Workbench for group and voluntary benefits.

Our read

Carriers on Majesco get agentic features inside the policy, claims and underwriting screens they already use, with governance controls aimed at NAIC expectations. Majesco names no customers or underlying models, and its claim of workbooks built in as little as 30 minutes has no stated method.

Model ML

Oct 6, 2026

ImpactMedium

VerificationVerified

Model ML's September 2026 update splits larger tasks across agents running in parallel by default, lets agents work across open Excel workbooks with before and after values and undo on every changed cell, and adds a native Slack integration in which the agent works in a thread with the same connectors as the web app. Bigdata.com news, expert interviews, transcripts and sentiment data are now live in the platform, the Document Review Agent can validate against call transcripts or use Excel files as the target, and Sites gains building from chat or email, viewer analytics and scheduled refreshes.

Our read

Banking and advisory teams can hand Model ML larger research and modeling jobs and work with it in Slack, with every spreadsheet change traceable. This is a monthly roundup, and some items, such as the Bigdata.com expansion, were announced earlier in September.

Owlin

Oct 6, 2026

ImpactMedium

VerificationVerified

Owlin's Q3 2026 product release makes event extraction fact based, building each event from concrete details such as dates and locations and no longer merging separate stories into one event, and makes risk scores context aware, weighting a development by its likely impact on the specific company given its size. Grouped alerts send one email listing every channel a development triggers instead of one per channel, and each alert channel can now export its alerts for a chosen period as a spreadsheet.

Our read

Credit and third party risk teams monitoring large counterparty lists get fewer duplicate alerts and scores that reflect the company in question, and can pull alert records into their own review and audit files. Owlin gives no metrics, and the Q3 features may have reached early clients before the post went up.

Persona

Oct 6, 2026

ImpactMedium

VerificationPartially Verified

Persona made Vision Mode for document extractions generally available. Instead of reading text with OCR and extracting from it, Vision Mode passes the document image straight to the AI model so it considers layout as well as text, aimed at non English documents and complex layouts such as Chinese business licenses. It is configured per extraction type, is part of the Enhanced Document Verification SKU enabled through the account team, and adds about 2 to 5 seconds to median processing time.

Our read

Teams verifying businesses and individuals across many countries can raise extraction accuracy on documents OCR handles badly, at the cost of a few seconds per document. Persona names no model and gives no accuracy figures.

Polly

Oct 6, 2026

ImpactHigh

VerificationVerified

Polly launched its next generation Hedge system, built into the PollyOS core rather than connected to it, completing a single loan level record for mortgage capital markets alongside its pricing engine, the reintroduced Loan Trading Exchange and secondary accounting. Hedge positions update from live lock, execution, LOS and market data, so cancellations, status changes and intraday fundings appear immediately and positions are shown against live pricing rather than a prior day snapshot, with margin attributed loan by loan using the same calculations accounting relies on.

Our read

Secondary desks can see their real hedge position and loan level margin during the day and act on market moves without reconciling several systems. Only an initial group of customers has it now, with broader availability after the MBA Annual Convention on 11 to 14 October.

Sapiens

Oct 6, 2026

ImpactHigh

VerificationVerified

Sapiens launched Sapiens Brain, the AI foundation for its SapiensAIP platform: an insurance native knowledge graph built from the code, product documentation and design artifacts of more than 40 years of Sapiens systems. SapiensAIP agents for claims, underwriting, migration and configuration run outside the core systems and reach them through MCP, so AI can update faster than the core release cycle, and Sapiens uses domain tuned models for contained tasks rather than one frontier model across a workflow, with agent decisions and reasoning logged for audit.

Our read

Insurers on Sapiens cores get agents grounded in how their own systems are built, with an audit log of agent reasoning, which matters for model governance. Sapiens' claim that it eliminates hallucinations has no published evaluation, and MiWay is the only named customer.

LendAPI

Oct 6, 2026

ImpactMedium

VerificationVerified

LendAPI 3.2.9 adds two webhook steps to its Flow Builder. Send webhook pushes chosen application data to an outside system, and Wait for webhook pauses the application until that system replies, with a maximum wait the lender sets and a separate path if time runs out. The release also adds an undo for every builder on the platform.

Our read

Lenders can keep outside manual reviews, partner approvals and returned documents inside one origination flow instead of chasing them by hand, and the timeout path stops applications stalling when a partner system does not answer.

Vikar Technologies

Oct 6, 2026

ImpactMedium

VerificationPartially Verified

Vikar added FIS ChexSystems risk screening to its account opening platform, so community banks and credit unions can run ChexSystems, QualiFile and BizChex checks inside Vikar for both consumer and business accounts.

Our read

Institutions that already screen applicants through ChexSystems can keep that deposit risk check inside their account opening workflow without a separate system or rekeying, and BizChex extends it to business accounts.

Chronograph

Oct 5, 2026

ImpactMedium

VerificationVerified

Chronograph released two plugins for Claude, Chronograph GP and Chronograph LP, each pairing its existing MCP connector with a curated set of skills for Claude Cowork, Chat and Code. GP skills assemble quarterly review packs, generate branded investor reports from live Chronograph data, compare actuals with budget, brief valuation changes ranked by NAV impact and attribute fund TVPI by company. LP skills forecast capital calls, distributions and NAV with the Takahashi-Alexander model, build commitment pacing plans, show look through exposure and prepare questions for manager meetings.

Our read

Private equity GPs and LP allocators can run multi step reporting and forecasting in Claude on validated portfolio data without building their own pipelines or prompts. The plugins are available to all Chronograph clients; no results or customer names are given.

Hebbia

Oct 5, 2026

ImpactMedium

VerificationVerified

Hebbia's October 2026 release adds memory to Max, its agent, which stores and recalls context across conversations under user controls and an org level admin switch, version history with rollback for decks and docs made in Max, and a Microsoft Word plugin for running workflows and drafting inside Word. Matrix columns can now run conditionally on a prior column's output, ICE market data extends to Tokyo Stock Exchange and ASX tickers, and Intralinks data room content and Dynamo pipeline documents now sync into Hebbia.

Our read

Deal and research teams can draft in Word and pull data room and CRM documents into analysis without exporting files. Memory stays off until an admin enables it, which matters for firms with strict information barriers.

LeapXpert

Oct 5, 2026

ImpactMedium

VerificationVerified

LeapXpert released an integration that lets Webex users message external clients on WhatsApp from inside Webex, available now on the Webex App Hub. Users sign in with their Webex credentials on desktop or mobile, messages pass data loss prevention checks in both directions, and LeapXpert captures the full conversation for archiving. It requires a Leap Work account and a Webex license, and WhatsApp is the only channel at launch.

Our read

Banks and brokers on Webex can meet off channel communication rules for WhatsApp without a separate messaging app for advisers. Further channels are planned with no dates.

Vega Minds

Oct 5, 2026

ImpactMedium

VerificationVerified

Vega Minds' October 2026 update adds a Claude connector, listed in Claude's directory, that gives read only access to clients, households, meetings, transcripts, documents and the Knowledge Center and stays off until a firm admin enables it. Vega Pro firms gain living stories, a daily updated profile of each client built from notes, emails and CRM data with every fact dated, compliance review of full meeting transcripts, and planning deliverables such as Roth conversion and RMD projections exported to PDF, Word or Google Docs. Vega names its models: GPT-6.1 Sol for client deliverables, Claude Opus 5.5 and Sonnet 5.5 for meeting prep and writing, and Gemini 3.8 Flash and GPT-6 Luna for background tasks.

Our read

Advisory firms can use Claude on their Vega data without giving it write access, and compliance can review what was said in meetings, not just what was sent. Several features are Pro only, ChatGPT support is still coming, and the post covers a month of releases, some of which predate the window.

WealthAi

Oct 2, 2026

ImpactMedium

VerificationVerified

WealthAi launched a suite of AI agents that run wealth client onboarding and KYC end to end. The agents collect client information and documents, identify what is missing, generate forms, and start ID verification, PEP and sanctions screening, KYC and account checks at the right stages, connecting to the verification services firms already use and writing straight into the WealthAi Client File and existing internal systems; human review comes in only where the firm's process requires it. WealthAi says the agents handle clients with interests across several jurisdictions.

Our read

Wealth managers whose onboarding stalls on chasing documents and rekeying data can hand that coordination to agents while keeping approvals where their policy demands. WealthAi names no customers, availability terms or time savings.

Hebbia

Oct 1, 2026

ImpactHigh

VerificationVerified

Hebbia introduced Headless Hebbia, an API and MCP server that brings its retrieval, orchestration and analysis into firms' own applications and into assistants such as Claude and ChatGPT. Retrieval and analysis run step by step so developers can test a workflow and users can check findings against source material, the system maps relationships across sources such as subsidiaries and differently named filings, and it supplies evidence to each model call so applications can work over large collections with smaller context windows.

Our read

Banks and investment firms that want Hebbia's document analysis inside their own origination or research tools no longer have to work in Hebbia's interface. Access runs through Hebbia's sales team, and no public API documentation was published with the launch.

Kore.ai

Oct 1, 2026

ImpactMedium

VerificationVerified

Version 1.7.2 of Kore.ai's Agent Platform lets builders choose how much of each tool result the model keeps in conversation history (the full result, selected fields or a template summary) and treats workflow and MCP tools as having side effects by default, so retries do not repeat operations and confirmation policies cover them. It adds an Oracle Database connector for queries and row changes, optional encryption of payloads returned by the Sessions, Conversations, LLM Logs and Traces APIs, filtered bulk delete for Agent Tables up to 10,000 rows, and an FAQ connector for knowledge bases. The idempotency header on HTTP tools is now opt in, and the voice guardrails introduced in 1.7.0 are now off by default.

Our read

Banks building agents on Kore.ai get protection against duplicate transactions on retries, encryption for exported conversation data and a direct connection to Oracle databases. Teams that relied on the 1.7.0 voice guardrails need to switch them back on for each agent.

Moody's Analytics

Oct 1, 2026

ImpactHigh

VerificationVerified

Moody's Analytics and Allvue launched the EDF-X Private Credit Model, which Moody's describes as its first model calibrated directly on observed private credit performance, using anonymized borrower data contributed by Allvue under contract so no individual firm can be identified. The model separately estimates the likelihood of hard credit events such as missed payments and of the soft events that often come first, such as covenant waivers and payment in kind arrangements. It is available now to customers of both firms through the Moody's Analytics EDF-X API.

Our read

Private credit lenders and investors get an early warning signal for unrated borrowers calibrated on the asset class itself rather than on public company proxies, for monitoring, manager evaluation and allocation. Moody's publishes no accuracy, backtest or validation results with the launch.

Ondato

Oct 1, 2026

ImpactLow

VerificationPartially Verified

Ondato made European Digital Identity Wallet verification available in the web version of its identity verification flow on 1 October 2026, according to AMLR guidance it published the same day. Support in the mobile app is due by the end of Q4 2026, alongside a qualified trust service provider route under eIDAS.

Our read

Institutions preparing for the EU AML Regulation, which applies from July 2027, can start accepting EU digital identity wallets in web onboarding. Ondato has not published product documentation, so which national wallets are accepted and how the flow works are not yet stated.

RightCapital

Oct 1, 2026

ImpactMedium

VerificationVerified

RightCapital's Q3 2026 release roundup expands Iris, its AI planning agent, with a Find Opportunities preset that reviews a client plan and flags gaps across tax, investments, retirement, insurance, debt, stock plans, business and estate, plus document upload so advisers can ask about annuity illustrations, stock grants or policies and compare a new file with one already in the client's Vault. The Medicare module was rebuilt as Health Care with cost analysis and side by side Medicare options, and planning gains exact year of death modeling for life insurance and the new federal Repayment Assistance Plan for student loans.

Our read

Advisers can cut meeting prep by letting Iris surface planning opportunities and read complex client documents. Iris is limited to Premium and Platinum subscriptions, and the roundup gives no per item dates, so some features went live earlier in the quarter.

Unit21

Oct 1, 2026

ImpactMedium

VerificationVerified

Unit21's September release adds configurable SAR autofill: teams map values once and later filings pull in FinCEN advisory codes, entity and instrument data, ID values, the filing institution's TINs and account numbers, and check items automatically from tags. Two new tools on Unit21's MCP server let an AI assistant list a report's filters and their valid values so it can build correct report queries, and post account warnings can now be sent to Jack Henry Symitar directly from Unit21 through the PortX partnership, by rule or by hand.

Our read

BSA and fraud teams spend less time retyping the same data into each SAR, and credit unions on Symitar can act on a finding in the core without switching systems. Unit21 calls this the first wave of autofill, with more coverage to come.

Root Platform

Oct 1, 2026

ImpactMedium

VerificationPartially Verified

Root's September release notes, published 1 October, extend collection modules to payment partners that exchange files instead of API calls. The platform sends the files, receives partner replies and settles the payments, with safeguards so each file is sent only once. Payment methods can also now be found and corrected by external reference and listed by collection module.

Our read

Insurers and distribution partners on Root can connect payment and debit partners that only accept batch files without building their own file transport and settlement logic.

CRIF

Sep 30, 2026

ImpactMedium

VerificationVerified

CRIF launched CRIF ORCHESTRA, a composable credit lifecycle platform for banks and lenders covering onboarding, decisioning, origination, monitoring and collections. Banks assemble reusable, API enabled business capabilities one use case at a time alongside existing systems, connecting data, models, rules and workflows in one governed environment, and can test and compare decision strategies before a wider rollout. Its decision intelligence module, CRESCENDO, adds AI augmented decisioning with explainability and human oversight, and Raiffeisen Bank Romania is the named user.

Our read

Lenders replacing a monolithic decision engine can modernize in stages and keep a traceable record of which rules and models drove each decision. The launch was a webinar with demo bookings, no availability terms are stated, and parts look like a repackaging of CRIF's existing StrategyOne platform.

Jump

Sep 30, 2026

ImpactMedium

VerificationVerified

Jump's September 2026 release adds a Genesys Cloud integration that ingests completed calls into the adviser's meeting record, a TradePMR integration that pulls live accounts, balances, positions and tax lots from the Fusion platform into meeting prep, and a deeper Black Diamond pull covering cost basis, unrealized gains, client timeline posts and portfolio groupings. Enterprise API endpoints now expose SCIM user metadata and contact record IDs, and firms can require a pre meeting prep attestation that creates an audit trail. The post also lists real time Schwab and Fidelity account opening, announced in limited beta on 14 September, which needs the Onboard upgrade and broker dealer approval.

Our read

Wealth firms get phone calls captured as meetings and fresher custodian data in prep without manual entry, and IT teams can reconcile Jump users and contacts against their own identity and CRM systems. Several items carry a caveat that they may not be available at every firm.

Ocrolus

Sep 30, 2026

ImpactMedium

VerificationVerified

Ocrolus's October 2026 release notes add a Conditions API for lenders not on Encompass, with endpoints to list, create, update and clear conditions on a Book and two webhook events for generated and updated conditions, plus a Doc Expiration tab that checks loan file documents against Fannie Mae and Freddie Mac age limits measured to the closing date. New endpoints generate and download the income analysis PDF as a ZIP and upload MISMO XML to an existing Book, Docs to Digital can now run on a Plaid file already in a Book, and Classify and Capture adds timestamp fields to Closing Disclosure and Loan Estimate forms.

Our read

Mortgage lenders outside Encompass get their first programmatic route to Ocrolus conditions, and expired documents surface before closing instead of at audit. Doc Expiration needs its own entitlement plus an Automated Conditions subscription and covers conventional and GSE loans only.

Shortcut

Sep 30, 2026

ImpactMedium

VerificationPartially Verified

Shortcut release 1.1.0 added Claude Sonnet 5.5, GPT-6 Sol and GPT-6 Luna to its model picker and made background tasks follow the user's chosen model by default. The release also added earnings forecast pages for exploring standalone financial projections, plus admin changes that verify SSO domain changes before applying them and simplify SCIM onboarding.

Our read

Finance teams with a model governance policy can pick from more frontier models and keep background work on the approved one. The lineup moves quickly: five days later GPT-6.1 Sol replaced GPT-6 Sol in the picker.

Blend

Sep 30, 2026

ImpactHigh

VerificationVerified

Blend made Navigator, the Autopilot agent loan officers chat with inside the loan file, generally available. Each approval is now bound to the exact values Navigator showed on the approval card, and if the loan changes after the card is created, Navigator holds the action and surfaces the conflict instead of acting on information the loan officer did not review. An approval given on one device resolves everywhere.

Our read

An agent that can act on a live loan file now cannot carry out a change that differs from what the loan officer approved, the kind of control a lender's compliance and model risk teams will want to test.

Akur8

Sep 29, 2026

ImpactHigh

VerificationVerified

Akur8 launched Akur8 Agents, starting with one agent for each module of its insurance pricing platform. The agents build pricing structures, write formulas, generate reports from reusable templates, check variable usage before modeling, apply filed rates in rating analyses and bring rating logic from spreadsheets into Akur8 through a reviewable process, and teams can add expert skill templates or build custom skills without code. Akur8's own actuarial engines still do the calculations, users inspect the reasoning and approve changes before they run, and customer data is not used to train AI models.

Our read

Pricing and actuarial teams get help with the slow parts of model building and reporting while the math stays in Akur8's engines and every change needs approval, which matters for rate filing and model review. Akur8 names no customer using the agents yet, and agents for reserving and life are planned, not shipped.

Arch

Sep 29, 2026

ImpactHigh

VerificationVerified

Arch launched Arch Investment Research, extending its private markets platform from post commitment work into pre investment due diligence. It extracts key terms from offering materials, links each term to its source document and lets investment teams evaluate opportunities against their own criteria, and once a deal closes the diligence materials carry into ongoing monitoring. It is sold standalone or with Arch's existing data management, capital call and reporting suite, aimed at RIAs, family offices and institutional allocators.

Our read

Allocators reviewing a steady flow of private market offerings can cut manual term extraction and keep a source linked record from diligence through monitoring. The time savings Arch cites (up to 50% per opportunity, two hours on one Clearstead deal) are customer anecdotes without a method, and analysis of Arch data through customers' own LLMs is planned, not shipped.

Bretton AI

Sep 29, 2026

ImpactMedium

VerificationPartially Verified

Bretton improved Agent Builder so teams can build, edit and validate the structured results and text reports their agents return, with guided field editing, setup from a sample, and validation before a change is saved. Changing a downstream output format previously meant manual configuration and test runs.

Our read

Compliance teams feeding agent output into case systems, exports or APIs can change the output shape themselves with less risk of breaking an integration. The changelog tags this as an improvement to an existing capability and gives no plan or rollout detail.

Finastra

Sep 29, 2026

ImpactMedium

VerificationVerified

Finastra announced Repair Recommendations, an AI capability in AI OperatorAssist that identifies payment discrepancies, analyzes the cause and proposes fixes validated against scheme and network rules for Swift, Fedwire, SEPA, UPI and Nexus. Every recommended action goes to a person for review and approval, and those decisions feed back into the recommendations. It works inside existing repair workflows and complements Global PAYplus and Payments To Go.

Our read

Payment operations teams spend much of their exception handling time working out why a message failed, so scheme aware suggestions with mandatory human approval target a real cost without handing over control. The release calls it a launch but gives no availability date, version or measured result.

Hawk

Sep 29, 2026

ImpactHigh

VerificationVerified

Hawk added a Customer Due Diligence Agent to its agentic AI platform. It runs on an institution's own approved policies, combines KYC profile changes, screening hits and transaction alerts into one assessment, cross checks declared profiles against actual transaction behavior, maps multi layer ownership chains from customer data, corporate registries and adverse media, and hands analysts a prefilled CDD or EDD review with narratives, explanations and clickable citations, triggering enhanced due diligence when risk thresholds are crossed. The agent is LLM agnostic and never closes, escalates or submits a case without human approval.

Our read

AML teams buried in periodic reviews can move evidence gathering and form filling to an agent while analysts keep every decision, which is the oversight regulators expect. Hawk names no customers and gives no time savings or GA status.

Persona

Sep 29, 2026

ImpactMedium

VerificationVerified

Persona launched two managed fraud products. Escalated Verifications runs its most compute heavy fraud models, including multi frame video analysis and clustering models that need data gathered after onboarding, once a Government ID or Selfie verification is done, then marks medium or high risk verifications as escalated with a dashboard warning and a webhook that can trigger cases, extra checks or account suspension. Fraud Fight Flow is a Persona maintained verification flow, kept current from Persona's research and the attacks seen in the customer's environment, that returns a pass or fail and can sit before verification, after data collection or as a step up.

Our read

Fraud teams can apply heavier detection without adding friction at onboarding, and hand upkeep of an anti fraud flow to Persona instead of their own engineers. Escalated Verifications is limited to Enterprise plans, Fraud Fight Flow is early access, and both were offered in early access over the summer before this launch post.

Persona

Sep 29, 2026

ImpactMedium

VerificationPartially Verified

Persona added AI assisted fraud checks for supplemental documents, the files other than government IDs that customers upload during verification, in early access. Customers set their own risk definitions for dedicated checks on AI generated documents, visual tampering, PDF tampering, low effort fabrication and suspicious content, a Fraud Risk Assessment check summarizes the results, and settings can differ by document verification template.

Our read

Generated and edited supporting documents are a fast growing onboarding fraud route, and per category checks let a risk team tune what counts as high risk rather than accept a single score. It is early access, and the underlying document fraud detection sits on Growth and Enterprise plans.

Persona

Sep 29, 2026

ImpactMedium

VerificationVerified

Persona released API version 2026-09-29. Inquiry Sessions gain a gps-country-code attribute, an ISO country code derived by reverse geocoding the session's GPS coordinates and returned in session endpoints and webhooks, and redacting a session now also clears its GPS coordinates, geocoded location and stored webhook copies. The version carries three breaking changes for customers who upgrade: UK Share Code immigration status returns the Home Office's own text, webhook creation requires a name, and Business Adverse Media related sources return only documented fields.

Our read

Compliance teams get a cleaner geographic signal for each verification session and stronger redaction for privacy requests, and developers can see exactly what breaks before moving to the new version. Breaking changes apply only when a customer upgrades its API version.

WTW

Sep 29, 2026

ImpactMedium

VerificationVerified

WTW released Radar AI Assistant, a natural language capability inside Radar Vision, its performance and experience monitoring tool for insurers. Pricing, underwriting, claims and portfolio management teams can ask questions of their own data to spot emerging issues and patterns, and the assistant combines weak early indicators into one view, explains the likely drivers and recommends pricing and underwriting actions.

Our read

Insurers already on Radar, which WTW says more than 500 use, get faster portfolio diagnosis inside the tool they run rather than another reporting layer. WTW states no availability terms, customers or results for the assistant.

ACI Worldwide

Sep 28, 2026

ImpactMedium

VerificationVerified

ACI Worldwide and Cognizant published results from a performance test of BASE24-eps, ACI's card acquiring and issuing switch, running on AWS managed services (Amazon EC2, Amazon RDS for PostgreSQL, OpenTelemetry and Amazon CloudWatch). In a non production environment under simulated point of sale load, the switch sustained more than double its target transaction load for 30 minutes and processed more than 4.6 million transactions with zero failures. The release does not give the target figure, throughput or latency, and says production use still needs failover and disaster recovery testing, security hardening and payment network certification.

Our read

Issuers and acquirers weighing a move of an existing BASE24-eps estate to AWS get a stated test design and a failure count to plan against, without a rebuild on a new platform. It is a vendor run test with no bank named and no throughput figure, so treat it as a starting benchmark rather than production evidence.

Duck Creek Technologies

Sep 28, 2026

ImpactHigh

VerificationVerified

Duck Creek launched Agentic FNOL, an agent based first notice of loss product built on the Duck Creek Agentic AI Platform with Google Cloud and Gemini models, now available to early access customers. Claimants describe a loss in their own words across digital, voice to text and mobile channels; the agents decide which follow up questions to ask, verify policy and coverage, assess injury and legal severity, flag anomalies against submitted images, policy data and the wider claim record, and return a structured claim file. Triage and routing follow carrier defined rules, every AI determination is logged, and exceptions go to human review.

Our read

Carriers on Duck Creek Claims, or on another core system, can test conversational claims intake that hands adjusters a validated, triaged file instead of a form to clean up. Access is early access only, and Duck Creek gives no customer or cycle time results yet.

Fenergo

Sep 28, 2026

ImpactMedium

VerificationVerified

Fenergo's Agency Bulk Upload no longer requires a Managed Relationship file: during load, the platform finds the relationship between each Investment Manager and Underlying Principal, reuses any that exist and creates draft relationships where none do. A new Autoname setting, on by default, names each relationship from the two legal entity names instead of the generic label every system created relationship carried before, and a companion setting does the same for ETL Agency Migration projects. The change is live on non live tenants now and reaches live tenants from 5 October.

Our read

Banks and asset servicers onboarding agency structures with many underlying principals skip a separate relationship file and get records they can tell apart, which cuts data preparation and review time. It is an operational improvement to onboarding data handling rather than an AI capability.

Finastra

Sep 28, 2026

ImpactHigh

VerificationVerified

Finastra launched Finastra Supply Chain Finance at Sibos 2026, a cloud native module covering the supply chain finance lifecycle from buyer and supplier onboarding through fulfillment, servicing, risk management and native secondary market distribution with automated pricing, limit management and reconciliation. AI is embedded for fraud detection, compliance screening and eligibility monitoring, and the Nexus API suite links the module to Trade Innovation, Loan IQ and corporate ERP systems so limits and exposures sit in one view across trade finance, loans and supply chain finance. Purchase order, pre and post shipment, inventory and distributor finance are listed for future releases.

Our read

Banks already running Loan IQ or Trade Innovation can add supply chain finance without a separate stack and distribute risk to the secondary market from the same module. No bank is named as live and several finance types are still roadmap, so confirm scope in a demo.

FIS

Sep 28, 2026

ImpactMedium

VerificationVerified

FIS integrated Spade's real time transaction enrichment into its debit processing platform, available to FIS debit issuing clients now. Each card transaction is matched to an enriched merchant identity, so issuers can show clearer merchant names, locations and transaction details to cardholders and use the same data in authorization decisions, dispute management, analytics and AI features. Spade says its merchant data covers 99.9% of US and Canadian businesses.

Our read

Debit issuers on FIS can cut disputes caused by unrecognized charge descriptions and feed merchant data into fraud and authorization rules without a separate vendor integration. FIS gives no pricing, opt in details or measured dispute reduction.

InvestSuite

Sep 28, 2026

ImpactMedium

VerificationVerified

InvestSuite launched Charlie for Advice, an AI agent that banks embed in their own investment app or website to give retail customers personal investment advice. The customer states a goal in plain language, Charlie asks follow up questions, explains the options and builds a proposal with pessimistic, realistic and optimistic outcomes, then stays on to answer questions and rebalance when the portfolio drifts. The bank sets which data, investments and actions Charlie may use and when a human adviser takes over, and MiFID II suitability checks still apply.

Our read

Banks that cannot staff advice for customers with smaller sums get an advisory agent with bank set limits and a human handoff, a step beyond the April version of Charlie, which explained but did not advise. InvestSuite says it is available to banks today, but its pilot banks are unnamed and the first live customers are expected in 2027.

Jumio

Sep 28, 2026

ImpactMedium

VerificationVerified

Jumio completed the worldwide rollout of selfie.DONE, its reusable identity product, with its launch in EMEA after Latin America, North America and Asia Pacific. Previously verified users take a selfie first route instead of rescanning an ID document while required compliance checks run in the background, and each reuse needs the user's consent and runs fresh biometric matching and risk analysis rather than relying on cached results. Jumio cites an identity graph of more than 75 million identities and frames the EMEA design around GDPR, PSD2 and AMLD.

Our read

Banks and fintechs in EMEA can cut onboarding friction for returning or cross platform customers without dropping biometric and risk checks on each reuse. Jumio's figures of up to 20% higher completion and up to 60% of returning users verifying without documents carry no stated method or named customer.

Model ML

Sep 28, 2026

ImpactMedium

VerificationVerified

Model ML made Anthropic's Claude Sonnet 5.5 available in its platform after testing it with Anthropic on the Model ML Composite, its own benchmark for AI in financial services. Model ML reports that Sonnet 5.5 stays close to Opus 5.5 in four of the five benchmark categories at 56% to 78% lower cost in all five, sits near the top slide builders for PowerPoint creation at less than half their cost, and trails open models such as DeepSeek 4.1 Flash and Kimi K3 on pure play financial workflows.

Our read

Teams choosing which model to run for decks, memos and multi document work get a stated cost and quality comparison by task type, which helps set model policy. The comparisons come from Model ML's own benchmark with no published scores, so treat them as vendor reported.

S&P Global Market Intelligence

Sep 28, 2026

ImpactMedium

VerificationVerified

S&P Global added RatingsXpress credit data from S&P Global Ratings to the S&P Global AI Data Portal, which serves S&P data into MCP enabled AI applications. The addition covers nearly 1 million issuer and security level credit ratings plus outlooks, CreditWatch status and rating actions, available now through Deterministic Retrieval, the portal's structured, API driven query mode. Adaptive Retrieval, which lets agents and LLMs pull across several S&P datasets in natural language, is due for RatingsXpress in the coming weeks.

Our read

Credit, portfolio monitoring and counterparty teams building AI workflows can ground them in authoritative ratings data through MCP rather than a separate feed. Natural language access to the ratings is not live yet, and pricing and historical depth are not stated.

Vertafore

Sep 28, 2026

ImpactMedium

VerificationVerified

Vertafore made its Velocity Document Intelligence Agent generally available in ImageRight, five days after releasing it to early adopters. Users ask questions in natural language and the agent searches the Word files, spreadsheets, PDFs, images and email they are authorized to see, matching on meaning rather than keywords, and returns one answer with document and page level citations, for tasks such as summarizing coverage changes between policy terms or rebuilding the communications around a claim. Hummel Group Insurance is the named customer.

Our read

Agencies and carriers on ImageRight can now put cited document search into client calls, renewals, submission review and audit prep without a separate tool. Vertafore's figure of about 20 minutes of searching cut to two comes from early testing with no stated method.

AI Rudder

Sep 26, 2026

ImpactMedium

VerificationPartially Verified

AI Rudder published self evaluated τ-voice benchmark results for its voice agent: 83.89% pass@1 with the Standard customer simulator (237 of 278 tasks) and 88.82% with the Custom simulator (250 of 278 tasks) across retail, airline and telecom, above the prior published best of 81.72% and 86.23%. The post discloses the evaluated stack (Deepgram nova-3 speech recognition, gemini-3.7-flash, ElevenLabs flash v2.5 speech, no training or fine tuning) and the runtime controls around the model, including an enabled check that blocks a repeated identical tool write within a turn.

Our read

Lenders running voice agents on collections and payment calls need evidence the agent completes account actions once and correctly, and this gives a stated method and a named model stack to test against. The scores are self reported, exclude the τ-voice Banking domain, and were not on the official leaderboard when checked.

Bureau

Sep 25, 2026

ImpactMedium

VerificationPartially Verified

Bureau introduced Adaptive Authentication, which rescores risk continuously through a customer session using identity graph, behavioral, device, and location and network signals instead of deciding once at login. When risk rises at moments such as beneficiary addition or a high value transfer, the platform can trigger step up authentication (MFA, face or biometric checks) or block the session, and session signals feed Bureau's Graph Identity Network.

Our read

Banks fighting account takeover and authorized push payment fraud get in session step up decisions from the same engine used for onboarding. The pilot figures (50 to 60% more high risk fraud detected, 5 to 6% fewer false positives, step up on 1 in 70 sessions) come from unnamed tier 1 banks with no stated method, so ask for the evaluation design.

CipherOwl

Sep 25, 2026

ImpactMedium

VerificationVerified

CipherOwl's SR3 CLI release 2026.4.0 regroups commands under screen, reason, research, report, case, auth, config and agent, and new agent commands describe each action's typed inputs, outputs, effects and retry rules so an AI agent can call them safely. Batch screening results now keep probability, confidence, evidence, behavior decisions and model metadata when the service supplies them, and partial or failed outcomes are reported explicitly with a nonzero exit status.

Our read

Compliance teams screening crypto addresses from scripts or agents get results that carry the model's confidence and evidence, which supports review and model risk documentation. Partial results can no longer pass silently as clean ones.

ibl.ai

Sep 25, 2026

ImpactHigh

VerificationVerified

ibl.ai extended its PII and PHI filtering to in chat file uploads, covering text, Office documents, images (OCR plus pixel redaction) and PDFs (rasterized, redacted and rebuilt) on every path where a file reaches an LLM, with per agent block, redact or allow settings. Privacy mode now also keeps conversation text out of run tracking, the moderation log, application logs, Sentry events and Langfuse traces. A new read only privacy flags API records every PII or PHI detection across uploads, prompts, responses and memory retrieval, storing only the entity type, never the raw value.

Our read

Bank and insurer staff routinely upload statements and applications that carry SSNs and account data, so redaction before any model sees the file, and keeping chat text out of third party observability tools, narrows GLBA exposure.

ibl.ai

Sep 25, 2026

ImpactMedium

VerificationVerified

Admins can now define named, platform scoped network policies for agent sandbox VMs: a host and port allowlist sent to the sandbox each session, plus a custom egress profile per agent, and a missing or cross tenant policy fails closed. Sandbox secrets can also be drawn from stored integration credentials at runtime so rotations propagate, and sandbox runtime is metered at a configurable default of $1 per 10 minutes.

Our read

Agents that write and run code against financial data can be restricted to approved endpoints, which limits exfiltration risk and gives security teams a reviewable control before they allow autonomous runs.

Lendflow

Sep 25, 2026

ImpactMedium

VerificationVerified

Lendflow launched a hosted Model Context Protocol (MCP) server that connects a customer's Lendflow environment to AI assistants, available as a Lendflow plugin in ChatGPT and as a custom connector in Claude and Grok. Authenticated users can query applications, offers, funders and statuses in natural language and can also take write actions, including creating applications, adding notes and uploading documents, with the available actions set by the user's Lendflow permissions and configuration.

Our read

Lenders, brands and brokers can work loan files from the AI assistant their staff already use without building or hosting their own MCP infrastructure. Because the connector can create applications and upload documents, risk and operations teams should review which roles get that write access through an assistant.

Microblink

Sep 25, 2026

ImpactMedium

VerificationVerified

BlinkCard SDK releases published on GitHub on 25 Sep (iOS v3000.1.0, Android v3001.0.0, with Web 3001.0.0 on npm on 24 Sep) make redaction of sensitive card data the default. The card number is now redacted in both the captured image and the returned result with only the first 4 and last 4 digits visible, and the CVV is fully redacted; IBAN and cardholder name are not redacted.

Our read

Issuers, lenders and payments firms that capture cards in their apps now receive masked PAN and no CVV from scans unless they change the setting, which limits how much full card data lands in their systems and logs and bears directly on PCI DSS scope.

Model ML

Sep 25, 2026

ImpactMedium

VerificationVerified

Model ML added access to Blackroom AI virtual data rooms over MCP, so users can work with their Blackroom data room documents directly in the Model ML platform without downloading files or searching a separate data room. The connector is available now to anyone with an active Blackroom data room.

Our read

M&A and private equity deal teams can run Model ML diligence and drafting work against live data room contents instead of local copies, which keeps deal documents inside the data room. It only helps firms whose deals run on Blackroom.

Terminal X

Sep 25, 2026

ImpactMedium

VerificationVerified

Terminal X published its first Terminal X Bench scorecard for new frontier models, grading Opus 5.5, Sol 6 and Luna 6 separately on retrieval, processing and answer generation using medium thinking, the same test sets and the same grading prompts. Opus 5.5 led retrieval (80% average versus 56% for Sol 6 and 51% for Luna 6) and processing but was kept out of production because it refused routine finance prompts, and Sol 5.6 remains the answer generation model after its answers were preferred over Sol 6 in 68% of head to head comparisons.

Our read

Investment firms get a stage level record of which foundation models the platform tests, which one runs answer generation, and why a new release was held back, which supports third party model change control in a model risk program. It indicates new model releases are gated on per stage testing before routing changes.

Ruleguard

Sep 25, 2026

ImpactLow

VerificationVerified

Ruleguard 4.112.0 makes Regulatory Change Update tasks a measure of their own in MI 2. Teams can now report on those tasks directly by task owner, alert list, parent update and whether reminders are on, and filter them to the logged in user for a 'my open tasks' dashboard. The same release reworks MI loading for speed and adds a Refresh button to the dashboard toolbar.

Our read

Compliance teams at banks, insurers and asset managers can show who owns each regulatory change action, and whether it is overdue, straight from a dashboard instead of rebuilding that view from exported update records.

Socure

Sep 25, 2026

ImpactMedium

VerificationVerified

Socure's RiskOS platform is live as the identity verification, fraud prevention and risk decisioning layer on Arc, the Layer 1 blockchain built by Circle, at its public mainnet launch. RiskOS is integrated into the Arc Onramp flow that moves users from fiat into USDC inside any application built on the network.

Our read

Payments, stablecoin and digital asset teams building on Arc get a KYC and fraud decision step already wired into the fiat to USDC onramp instead of integrating one themselves. It extends RiskOS coverage into on chain money movement for crypto and fintech programs.

Baselayer

Sep 24, 2026

ImpactMedium

VerificationVerified

Baselayer replaced GET /autocomplete with a session based Autocomplete API: the backend mints a short lived session with POST /autocomplete/sessions and the browser calls GET /autocomplete/businesses on each keystroke, so the API key never reaches the client. Each suggestion now describes a canonical business with its officers, registered agents, addresses and liens, and carries a token valid for 15 minutes that is redeemed as business_token on POST /searches. A JavaScript/TypeScript SDK with a ready made typeahead component handles session refresh and retries, and the old route was set to stop working on 25 September.

Our read

Onboarding flows that use business typeahead had one day of notice to migrate, and the new token ties the KYB search to exactly the business the applicant picked, which cuts mismatched entity searches. Keeping the API key server side also removes a credential exposure risk in client side integrations.

Daloopa

Sep 24, 2026

ImpactMedium

VerificationPartially Verified

Daloopa's MCP integration now delivers its verified, source linked fundamental data inside Model ML, an AI platform that drafts pitch decks, due diligence reports and investment memos for financial and professional services firms. Model ML users can run comps, benchmark peers and monitor portfolio positions as new filings land, drawing on Daloopa coverage of 6,000+ companies.

Our read

Bank, asset manager and advisory teams drafting deliverables in Model ML can tie AI generated figures back to the original filing, which cuts the checking work on analyst output. It adds another third party AI workspace where an existing Daloopa subscription works without custom integration.

Elliptic

Sep 24, 2026

ImpactLow

VerificationVerified

Elliptic launched Pulse, an AI tool that lets a frontline law enforcement or government officer enter a wallet address or transaction hash and get a plain language risk summary showing the triggered rules, detected entities, exposure and flow of funds. One click escalates the full analysis, reasoning and field notes, tagged with a case number, into Elliptic Lens for screening and Investigator for tracing. Pulse is available now to law enforcement and government agencies.

Our read

It shows Elliptic extending its attribution data into agentic, explainable summaries where the AI informs but does not make the decision, the same oversight pattern financial institutions should expect in its compliance agents. Bank buyers get no new feature here, but case handoffs from police using Pulse arrive already structured for Lens and Investigator.

Elysian

Sep 24, 2026

ImpactHigh

VerificationVerified

Elysian launched two AI claims products for insurers. Ely Audit reviews open and closed claim files across a full book instead of a roughly 2% sample, evaluating handling quality, adherence to standards and external vendor performance, and flags portfolio wide patterns. Ely Adjust reviews all open claims daily, risk ranks them and recommends next steps to adjusters, with explainable guidance and an optional supervisor review layer; both connect to the carrier's existing claim system and can be bought separately or bundled.

Our read

Carriers and TPAs can buy Elysian's claims AI as standalone software rather than only through its TPA service. Ely Adjust only recommends, and the adjuster makes every decision, so it fits claims teams that need human sign off on each file.

FE fundinfo

Sep 24, 2026

ImpactMedium

VerificationVerified

FE fundinfo connected Nexus for Financial Advisers to the intelliflo adviser CRM. Through the Nexus Assistant meeting workflow, advisers can now pull intelliflo client data into Nexus and push information captured in Nexus workflows back into intelliflo. FE fundinfo says further CRM integrations will follow by the end of 2026.

Our read

Advice firms that run intelliflo can use the Nexus AI meeting and planning workflow without rekeying client data, and the CRM stays the client record of truth. Earlier Nexus releases only described sending meeting outputs to a connected CRM, so the pull direction is the new part.

Feedzai

Sep 24, 2026

ImpactHigh

VerificationVerified

Feedzai launched Farol, an AI agent embedded in its RiskOps Studio, with four launch skills: Risk Strategy (analyzes rule sets, flags rules that generate noise without catching fraud and recommends thresholds), Investigations (retrieves and summarizes alert data), Knowledge (answers product questions from documentation in the interface) and SAR Drafting. Feedzai states Farol runs inside each institution's own environment so the data it surfaces stays in the customer's estate, and that agent outputs come with traceable reasoning for analyst review.

Our read

Fraud and AML teams on Feedzai get agent drafted SARs and rule threshold recommendations inside the existing case workflow, so BSA and model risk owners will need to decide how agent outputs are reviewed, logged and approved. The 20% faster alert handling and 12x faster SAR drafting figures carry no stated method, so buyers should ask for the measurement basis.

FinTech Studios

Sep 24, 2026

ImpactMedium

VerificationVerified

FinTech Studios announced general availability of its MCP connector, now listed in the official Claude and ChatGPT connector directories and addable to Claude for Financial Advisors. The hosted server at mcp.fintechstudios.com uses OAuth 2.1 sign in and exposes read only tools for cited news, market and regulatory search, trending entities, AI summaries, daily briefs, collections and insight channels, metered against FTS Studio credits with a free tier. The underlying npm package and MCP Registry entry were first published on 15 Sep 2026, so the in window change is the directory listing and GA.

Our read

Compliance, research and advisory teams can pull cited regulatory and market intelligence straight into the assistant they already use without handing an API key to the AI provider. Because the tools are read only, the connector cannot create, change or send anything in the user's account.

FIS

Sep 24, 2026

ImpactMedium

VerificationVerified

FIS released the latest version of the FIS Retirement Platform (FRP), rebuilt on a cloud native microservices architecture that analyzes plan data in real time. FIS states it is available now to retirement plan providers and third party administrators of all sizes, and positions it to support SECURE 2.0 plan designs such as mandatory automatic enrollment. Participant guidance, back office automation and plan health analytics are listed as roadmap items, not shipped features.

Our read

Recordkeepers and TPAs weighing a move off legacy hosted recordkeeping now have a generally available cloud native option from FIS. AI capabilities remain roadmap only, so buyers should not count them in an evaluation yet.

LexisNexis Risk Solutions

Sep 24, 2026

ImpactMedium

VerificationVerified

LexisNexis Risk Solutions integrated its Drive Metrics telematics scoring model into the IMS EdgeSDK, a smartphone SDK that U.S. auto insurers can embed in their own app or use through an IMS white label app. Driving data is scored on the device, producing Drive Metrics scores and attributes consistent with LexisNexis Telematics Solutions, without sending trip data to cloud processing.

Our read

Insurers without a usage based insurance program get a lower cost route to telematics underwriting, with the vendor claiming 50% to 80% lower UBI operating cost from removing cloud processing and storage fees. On device scoring also narrows how much consumer driving data the insurer has to collect and retain.

Model ML

Sep 24, 2026

ImpactMedium

VerificationVerified

Model ML added a Daloopa connector over MCP, so users with an active Daloopa license can pull Daloopa's standardized fundamental data on 6,000+ companies directly into the Model ML platform. Each Daloopa data point links back to the SEC filing, press release or investor presentation it came from, so figures used in models, pitch decks and memos can be checked against the original document.

Our read

Banks and investment firms that already license Daloopa can use it inside Model ML agent workflows without exporting data, and source linked figures make AI built deliverables easier to review. Access depends on the firm holding its own Daloopa license.

Oz Forensics

Sep 24, 2026

ImpactMedium

VerificationVerified

Oz API 6.7.0 deprecates webhook delivery of analysis results through the resolution_endpoint folder field and states the mechanism will be removed in API 7.0.0. The same release fixes a slowdown on GET /api/folders/ when resolution status and analysis type filters are combined with with_analyses=true.

Our read

Banks and fintechs that push liveness and face match verdicts into onboarding or KYC systems by webhook need to move to another result retrieval method before upgrading to 7.0.0.

Streetbeat

Sep 24, 2026

ImpactLow

VerificationVerified

Streetbeat shipped iOS app version 2026.9.0, which the release note calls a rebuild of the whole app. It adds The Beat, two AI written notes a day tailored to the user's holdings (an Opening Beat before the bell and a Closing Beat after), per stock, ETF and portfolio Discuss rooms that flag when a poster holds the asset, and an exposure view that counts holdings held through ETFs and portfolios. It also adds a portfolio library that can be invested in a Streetbeat account or a connected Robinhood account, configurable widgets, and support for seven languages.

Our read

Streetbeat says the consumer app runs on the same AI foundation it licenses to institutions as a white label AI Suite for Retail, so this release is a working sample of the holdings aware generated content and look through exposure analytics a bank or wealth manager would deploy under its own brand.

SymphonyAI

Sep 24, 2026

ImpactMedium

VerificationPartially Verified

SymphonyAI introduced Symphony Risk Intelligence (SRI), an agent native financial crime compliance platform built around a System of Intelligence that tracks changes in regulations, threats, internal policy and customer risk and directs agents to run detection, triage, investigation and reporting in response. Governance sets boundaries on what each agent may do, human oversight configurable from semi to fully autonomous, and a full audit trail, applied on the same terms to agents an institution builds itself. SRI layers over existing transaction monitoring and screening systems of record rather than replacing them.

Our read

Compliance teams can set how far agents act on alerts without an investigator and hold their own in house agents to the same audit trail, which bears directly on model risk review and examiner questions. The SRI acronym previously stood for Sensa Risk Intelligence, so buyers should ask which parts are new and generally available versus renamed.

Vertafore

Sep 24, 2026

ImpactMedium

VerificationPartially Verified

Vertafore introduced two Velocity AI agents for MGA underwriting: the Change Request Agent, which summarizes what changed across renewal and endorsement applications and requires underwriter approval before the summary is saved to the application record, and the Configuration Agent, which turns natural language admin requests into Surefyre program configuration that admins review, preview and test before saving. Both are listed as early adopter on the Velocity AI platform page, with general availability expected this fall.

Our read

MGAs and carriers underwriting through Surefyre get AI inside the underwriting record with an explicit human approval gate before anything is written. Buyers can hold Vertafore to that review before save control when assessing automated underwriting support.

Vise

Sep 24, 2026

ImpactMedium

VerificationVerified

Vise and Bitwise rolled out Bitwise's diversified crypto model portfolios to the wealth firms and 135,000+ accounts on the Vise platform, with Core models for broad crypto exposure and Thematic models for specific positioning such as risk managed exposure or assets beyond bitcoin. On Vise the models sit inside the same account as the client's equities, fixed income and alternatives, sized to the household risk profile, rebalanced on the same schedule and covered by the same tax management.

Our read

Advisors can offer a researched crypto allocation without running it as a separate book, so drift control and tax loss harvesting apply to crypto the same way they apply to the rest of the portfolio. This extends the 2025 bitcoin access to a menu of third party crypto models.

Ripjar

Sep 24, 2026

ImpactMedium

VerificationVerified

Ripjar upgraded the adverse media risk classifiers in its ULTRA engine, which read each news article to decide whether it relates to a specific risk and which person is involved, moving them to a new generation of machine learning model. The same update improves entity matching so screening can tell apart parent companies, subsidiaries and similarly named group entities, and handles more complex naming conventions, including Arabic names. Both feed Screening Assistant, the agentic triage feature that closes low risk alerts on its own.

Our read

A model change in the classifiers that decide adverse media risk is something model risk teams need to log and retest, and it changes which alerts the AI closes without an analyst. Ripjar named no model and published no accuracy figures, so buyers should ask for validation evidence on their own populations.

Sixfold

Sep 24, 2026

ImpactMedium

VerificationVerified

Sixfold launched Distribution Intelligence, which scores every broker on submission volume, quality, appetite fit and hit ratio using data from submissions already run through Sixfold. Underwriting leaders set targets such as submission completeness, hit ratio, segment growth or renewal rates, or accept targets Sixfold recommends, and the system then runs or recommends initiatives (quote more in appetite submissions from a broker, refer a case, reprioritize the queue) that flow into the AI Underwriter with a stated rationale. It is the first capability in a new Portfolio Intelligence Suite.

Our read

Carriers and MGAs get broker level performance steering inside the underwriting workflow rather than in a separate BI tool, and the agent now reorders and recommends work against portfolio targets, so buyers should review which initiatives run automatically versus only as recommendations.

S&P Global Market Intelligence

Sep 24, 2026

ImpactMedium

VerificationPartially Verified

Kensho released version 8.0.0 of kfinance, the Python library and proxy MCP server for the Kensho LLM ready API. The OAuth refresh token is now sent to /oauth2/refresh in a POST body instead of the URL query string, so it no longer appears in access logs or request logs, and the library replaces the requests and httpx clients with httpx2, which changes the exception types raised for HTTP and network errors.

Our read

Firms pulling S&P Global data into AI workflows through kfinance or the proxy MCP server stop exposing long lived credentials in logs by upgrading. It is a breaking change, so engineering teams must update error handling and make any custom AUTH_REFRESH_URL accept POST before upgrading.

Bloomberg

Sep 23, 2026

ImpactMedium

VerificationVerified

Bloomberg launched a Collateralized Loan Obligations (CLO) Data solution with normalized CLO reference data plus loan level data on every loan held in a CLO portfolio, covering deal structures, identifiers, payment information, coverage tests and tranche level performance indicators. It is delivered through Bloomberg Data License (SFTP, REST API or a client cloud environment), B-PIPE and Bloomberg Query Language (BQL), and aligns with the Terminal functions LLD<GO> and CPR<GO>.

Our read

Insurers, asset managers and banks holding CLO tranches can trace exposure from tranche to underlying loan to corporate borrower through FIGI mapping and join it with Bloomberg regulatory, risk and credit data. ASKB access to this data is only planned, so AI querying of CLO collateral has not shipped yet.

Entrust

Sep 23, 2026

ImpactMedium

VerificationVerified

Entrust IDV iOS SDK 100.15.0 now stores capture videos in a protected on device location and deletes them automatically once they are no longer needed. Hosted capture modules now run in a nonpersistent WebKit data store and SDK network requests no longer write to the on disk URL cache, so no verification session data is left in the host app's container. The release also adds an EntrustIdvTrial module to the Swift package and fixes NFC handling for all NFC capable documents.

Our read

Banks and lenders embedding identity verification in their mobile apps get less residual selfie, document video and session data on customer phones, which narrows the biometric and personal data exposure they must account for in privacy and GLBA reviews.

Glia

Sep 23, 2026

ImpactHigh

VerificationVerified

Glia launched Glia Relationship Management, a built in AI CRM that joins voice, digital and branch interaction history with core banking data such as balances, transactions, products and loans in one customer record. AI updates the record after every interaction, creates records for unmatched callers and chats, and puts a customer summary on the agent's screen when an interaction starts. The same release adds Glia Branch, which captures in person conversations and desktop activity for branch staff, feeds them into the same record, and gives staff real time AI guidance and governed agentic workflows.

Our read

Banks and credit unions get one view of a member across the contact center, digital channels and the branch without a separate CRM project. Branch staff get the same AI assistance as the contact center, inside the same governance.

Incognia

Sep 23, 2026

ImpactMedium

VerificationVerified

Incognia added behavioral biometrics to its web risk intelligence, so it can now assess mouse movement and clicks, keyboard interaction, and copy, paste and cut activity, including whether those events are trusted browser events or synthetically generated. The signals are evaluated alongside its browser, device, integrity, bot and automation, and Verified IP Location signals, with step up to Cross Device Authentication on a trusted phone when web evidence is not enough.

Our read

Banks and fintechs defending web logins and account changes get a signal for scripted or synthetic sessions without adding a challenge for every user. The post states the business, not Incognia, makes the final risk and authorization decision.

INSTANDA

Sep 23, 2026

ImpactLow

VerificationPartially Verified

INSTANDA's no code policy administration and distribution platform is now listed in Accelerant's Marketplace, giving MGAs and carriers in the Accelerant network direct access to it. INSTANDA Clear, its governed workspace for underwriting referrals, exceptions and complex adjustments with automatic logging of every action, is also available through the listing.

Our read

MGAs and carriers already in the Accelerant network get a vetted route to procure INSTANDA and Clear without a standalone vendor evaluation. The listing widens access to existing products rather than adding new function, and Accelerant's own listing page could not be opened to confirm it.

Juniper Square

Sep 23, 2026

ImpactMedium

VerificationVerified

Juniper Square launched Capital Intelligence, a new add on product inside its Investor CRM built on Nasdaq eVestment Market Lens data covering 30,000+ institutional firms, 70,000+ mandates and 100,000+ investor and consultant professionals. An AI matching engine proposes links between a GP's existing contacts and the Nasdaq dataset, a new Pipeline Builder feature searches and imports investor and consultant profiles by firm type, AUM and geography, and the system flags contact, consultant and mandate changes. Every AI match and record update stays a suggestion until a team member accepts it.

Our read

GPs get institutional LP and consultant data kept current inside the CRM without a separate data integration project, and firms that already license Market Lens can connect their existing subscription. The accept before write design keeps a person accountable for each record change the AI proposes.

Kore.ai

Sep 23, 2026

ImpactMedium

VerificationVerified

Kore.ai AI for Work v1.21.0 adds an Enable guardrail toggle to the Tools section of an MCP server, so account guardrails run before any tool executes. The release also gives administrators a Models tab in Workspace Settings to enable, disable and reorder the LLM models available to each workspace.

Our read

Banks and insurers letting employee agents call tools over MCP get a pre execution control point and can restrict which models each team uses, which supports model inventory and human oversight expectations.

Kount

Sep 23, 2026

ImpactMedium

VerificationVerified

Kount's 23 September release reworks Link Analysis in Payments Fraud, with color coded matching that shows which criteria each related order shares, address line 2 as a new linking criterion, and approve, decline, review and bulk decision actions directly from the table. Kount 360 also gains a Webhook History table listing sent webhooks by solution, event type and status, with retry for failed webhooks and test sends.

Our read

Fraud analysts can see why orders are linked and act on a whole ring from one screen. Integration teams can find and resend failed webhooks without submitting a real order.

Murex

Sep 23, 2026

ImpactMedium

VerificationVerified

Murex announced that its MX.3 capital markets platform has achieved certification on Google Cloud, so institutions can now deploy MX.3 in Google Cloud environments. Until now Murex's cloud page listed only AWS and Microsoft Azure as certified environments for running MX.3, which makes Google Cloud a third supported hyperscaler.

Our read

Banks and asset managers standardizing on Google Cloud, or running a multicloud setup for operational resilience and DORA style concentration risk, can now host their trading, treasury and risk platform there. The release says clients are exploring it but names no live client yet.

Pave

Sep 23, 2026

ImpactMedium

VerificationVerified

Pave added eleven consumer cashflow attributes that break gambling activity out by kind, covering sweepstakes casinos, sports betting, lotteries and prediction markets, instead of reporting a single gambling total. The new fields include sweepstakes casino payment amounts, counts, recency, share, income ratio and monthly trend, plus each subtype's ratio to total gambling and a count of gambling subtypes used.

Our read

Lenders can weigh sweepstakes casino and prediction market spend separately from lottery spend in their own risk and affordability models, rather than relying on one blended gambling figure.

Pegasystems

Sep 23, 2026

ImpactMedium

VerificationVerified

Pega released the 26.2 Association Compliance GA package for Smart Dispute for Issuers and Smart Dispute for Acquirers (versions '23 through 26.1) and for Smart Dispute Agentic Automation (25.1 and 26.1). It updates dispute questionnaires, chargeback qualification logic and Visa Resolve Online transaction inquiry elements to match Visa, Mastercard and American Express rule changes published April to August 2026, including removal of Amex ZIP Code Verification eligibility checks and India specific MCC exclusions. It also adds a new response parameter to the Mastercard Retrieve claim details and Retrieve clearing details API calls.

Our read

Card issuers and acquirers carry chargeback losses when dispute logic lags network mandates, so a vendor shipped rule package that back covers five releases cuts in house rule maintenance before scheme effective dates.

Vertafore

Sep 23, 2026

ImpactMedium

VerificationVerified

Vertafore released four new Velocity AI agents to early adopters: Document Intelligence Agent in ImageRight, Coach Agent inside AMS360 and Sagitta, Policy Intake Agent that extracts commercial policy data from source documents, writes it to the agency management system and creates an audit ready activity record, and Policy Check Agent that compares policy documents against AMS data and recommends corrections for user approval. General availability is scheduled between September and early fall 2026.

Our read

Insurance agencies get AI that writes policy data into their system of record, so buyers should check which steps run without review (intake) and which require approval (policy corrections). Time savings are vendor early adopter figures with no stated method.

Socotra

Sep 23, 2026

ImpactHigh

VerificationVerified

Socotra released Configuration Assistant, an AI agent inside Configuration Studio that creates, updates and visualizes insurance product configurations from natural language prompts, up to 3 uploaded requirement files (such as a filing) or an existing tenant configuration. It proposes a plan the user must confirm, shows the product as canvas and explorer views with plain language plugin code summaries, retries failed validation up to 5 times, and deploys to a new tenant or an existing test tenant.

Our read

Carriers can now build and test products on the core themselves instead of relying on the Agentic Configuration implementation service Socotra has offered since 2025. Plan confirmation and the test tenant only deploy limit keep a person in the loop before anything reaches production, and the feature is off by default until a Socotra representative enables it.

Solidus Labs

Sep 23, 2026

ImpactMedium

VerificationPartially Verified

Solidus Labs says HALO is now running on GTN's global order book, providing trade surveillance and transaction monitoring across equities, fixed income, derivatives, FX and digital assets. It also says GTN's 500+ institutional clients can get their own dedicated HALO tenants through GTN's brokerage infrastructure, so they do not have to buy and integrate a separate surveillance system.

Our read

Brokers, banks and asset managers that already run on GTN can get surveillance through infrastructure they already use, which may shorten onboarding. Solidus has not published documentation for the tenant setup, so buyers should confirm data separation and who operates alert review (GTN's team or their own).

SS&C Technologies

Sep 23, 2026

ImpactMedium

VerificationVerified

SS&C added an AI portfolio management assistant to Advent Genesis that lets investment teams find opportunities, pull portfolio insights and start workflows in plain language, and that can review positions, rebalance portfolios and harvest tax lots. SS&C executives have described it as in beta, with production targeted within a few months. The same release introduces ConnectIQ, an AI advisor engagement platform for the retirement market that combines participant data from several sources and flags outreach opportunities with predictive analytics.

Our read

An assistant that can rebalance and harvest tax lots puts action taking AI inside the portfolio workflow, so buyers should confirm what human approval applies before trades are generated. SS&C says the new tools run on its AI Gateway governance platform, which is the place to ask how participant data is governed.

AlphaSense

Sep 22, 2026

ImpactMedium

VerificationPartially Verified

AlphaSense added connectors for Snowflake and BigQuery, so users can ask questions of their firm's own data warehouse inside AlphaSense alongside external research. The Snowflake connector calls Snowflake's managed Cortex MCP server under each user's own role, and the BigQuery connector turns plain language questions into read only queries through Google's Conversational Analytics API. Both query data live under the user's own identity and do not ingest or store a copy.

Our read

Research and investment teams can combine internal portfolio or pipeline data with AlphaSense's external content in one answer. Because queries run live under each user's existing warehouse permissions, no second copy of the data needs to be governed.

Baselayer

Sep 22, 2026

ImpactMedium

VerificationPartially Verified

Baselayer launched its Agentic Identity Suite, extending its business verification and risk infrastructure to AI agents that transact on behalf of people and businesses. The suite groups Know Your Agent (verifying which platform runs an agent, which person or business delegated it, and whether that delegation can be proven), Counterparty Verification, and a Trusted Registry of verified businesses and authorized agents, alongside the MCP released on 8 September.

Our read

Banks and payment firms facing agent initiated account opening or payments get a vendor option for tying each agent action to a verified principal with an auditable delegation record. No public API documentation or pricing for Know Your Agent was found, so buyers should confirm what is generally available before planning around it.

Bretton AI

Sep 22, 2026

ImpactMedium

VerificationVerified

Bretton added configurable quality control sampling: administrators can now set disposition based QC sampling rules for each individual agent directly in Settings. Changing how often agent case outcomes are pulled for QC review previously required manual coordination with Bretton.

Our read

Compliance leaders can set how much human review each agent's alert dispositions receive, by outcome, and apply that policy consistently without engineering help, which is the oversight record examiners ask about when agents close AML and screening cases.

BriteCore

Sep 22, 2026

ImpactHigh

VerificationVerified

BriteCore released an FNOL Copilot that takes first notice of loss through conversation, validates policy coverage and creates the claim record in real time. It expanded its Submission Intake Copilot into a multi agent Submission and Quote Copilot that ingests applications, schedules of values and loss runs, flags missing data, applies rating and checks carrier appetite to produce a quotable policy. The same release adds an Underwriting Workbench inside the policy administration platform, where underwriting rules and system events create, prioritize and route work.

Our read

Carriers and MGAs on BriteCore get AI help at intake, quoting and claims without adding a separate vendor, and employees still approve the next steps the copilots recommend. The workbench puts underwriting decisions, work queues and policy communications in one place.

CLARA Analytics

Sep 22, 2026

ImpactHigh

VerificationVerified

CLARA launched Agentic Intelligence inside its CLARAty.ai platform: a set of AI agents that run on each complex casualty and workers' comp claim from first notice of loss to resolution, tracking changes in severity, litigation potential, fraud risk and closure opportunities and recommending next steps. The agents expose score lineage and versioning, data freshness checks, cohort drill downs, scoring rationale tied to specific claim documents and fields, and a conversational agent over structured and unstructured claim data, while adjusters keep the final decision.

Our read

The product moves from surfacing risk scores to recommending and helping carry out claim actions, so carriers and TPAs should confirm which actions still require adjuster sign off. The per score rationale and data lineage give claims audit and model risk teams an evidence trail they can review.

Cytora

Sep 22, 2026

ImpactMedium

VerificationPartially Verified

Cytora's 22 September 2026 release reports that Zurich Commercial Insurance's deployment of the Cytora agentic AI platform in mid market underwriting now spans more than 20 countries, with digitization accuracy raised from below 80% to 98% and manual submission triage time cut by 80%. The 98% accuracy and 20+ country figures are new: Cytora's 18 May 2026 Zurich case study reported 95%+ accuracy across seven countries at the foundation stage and 11 live countries, and had already cited the 80% triage reduction.

Our read

A named global insurer's figures give commercial insurers a reference point for extraction accuracy at multicountry, multilanguage scale. The release states no measurement method for the accuracy figure, so buyers should ask how it was measured.

DiligenceVault

Sep 22, 2026

ImpactMedium

VerificationPartially Verified

DiligenceVault published a client case study reporting that GEM Investments, an institutional asset allocator whose ODD team also acts as its compliance team, saves 1 to 3 hours per manager review using DV Assist for fund document analysis, turning meeting notes into drafts, and compiling first draft ODD memos inside the platform. The case study says GEM began its rollout in Q4 2025 and names confidence scoring and source citations on every extracted insight as the features that drove adoption.

Our read

A named allocator with a quantified time saving on AI assisted ODD and compliance memo work gives buyers a benchmark to test against in their own evaluation. The figure is vendor published, and the stated range depends on document volume rather than a described measurement method.

FactSet

Sep 22, 2026

ImpactMedium

VerificationVerified

FactSet acquired BCC Group International (bccg), a Germany based developer of cloud native and hybrid market data distribution software, and is bringing bccg's ONE Platform into its Real-Time Solution Suite. ONE Platform ingests, normalizes, entitles and distributes real time and static market data through a single API, with centralized entitlements, compliance reporting and cost controls, and can be deployed in cloud, on premises or hybrid models or run as a managed service. A FactSet ONE Platform product page is now live in the FactSet Marketplace.

Our read

Banks, trading firms and asset managers that run their own market data infrastructure can now buy both the feeds and the distribution and entitlement layer from one vendor, with a choice of deployment model. Existing bccg clients should confirm contract, support and roadmap continuity under FactSet ownership.

FactSet

Sep 22, 2026

ImpactMedium

VerificationVerified

FactSet added InSync Analytics' Excel based company models, which combine InSync reported Actuals and Guidance with FactSet broker consensus estimates, to the FactSet Workstation Company Snapshot report and StreetAccount News stories. The first 1,000 of more than 2,500 planned models are released, with weekly coverage expansion, segment level KPI consensus and Actuals updated within two hours of earnings. InSync's WEAVE Excel AI agent and a joined InSync and FactSet MCP server are announced as upcoming, not shipped.

Our read

Buy side research teams at asset managers and hedge funds get filing linked, prebuilt models inside a workstation they already license, which cuts model maintenance during earnings season. The AI agent and MCP pieces are future items, so buyers should not count them as available yet.

Fiddler AI

Sep 22, 2026

ImpactMedium

VerificationVerified

Fiddler Release 26.20 made a new Fiddler hosted model, Ministral 3 8B (fiddler/ministral3-8b), generally available for prebuilt LLM as a Judge and Custom Judge evaluators, with Fiddler stating it gives more accurate judgments and reasoning at a runtime comparable to its other hosted models. The same release refreshed the LLM Gateway catalog (adding gpt-5.6-luna and gemini-3.8-flash) and removed gpt-5.6, claude-opus-5 and claude-fable-5 from the built in list across OpenAI, Anthropic, Vertex AI and Azure AI.

Our read

Model risk teams that use Fiddler judge scores as monitoring evidence now have a named, versioned judge model to record in model documentation. Evaluators pointed at the removed gateway models must be repointed or added back as custom models, which affects validation continuity.

Marloo

Sep 22, 2026

ImpactMedium

VerificationVerified

Marloo launched the Knowledge tab, a three layer knowledge base that shapes every meeting note, advice document and Ask Marloo answer it produces. A read only Regional layer maintained by Marloo holds regulatory, tax and rate references (FINRA Rules, CFP Board standards and FOMC rates in the US, the FCA Handbook in the UK), a Company layer lets firm admins set firm wide instructions and upload advice frameworks, compliance documents, approved wordings and disclaimers, and a User layer holds each adviser's own style within those limits. The help center says it is rolling out gradually as a beta and is included on Plus plans and above with unlimited company knowledge documents.

Our read

Advice firms can write their compliance and house standards down once and have them applied when drafts are generated, with only admins able to change company rules, which shifts consistency checks from after the fact file sampling toward prevention. Buyers should confirm beta availability for their account before relying on it.

motif

Sep 22, 2026

ImpactMedium

VerificationVerified

motif's Clarity evaluation docs added a dated 22 Sep 2026 validation run for its private markets desk, the private-markets-analyst agent that underwrites one private company from the holder's documents and independent reporting. On a fixed 5 question corpus it scored judge 80 and 100% citation grounding, but reviewer sign off came on only 1 of 5 answers, so motif states the desk does not yet pass its release gate. The run cost $0.015 per answer at 83% cache and p50 112 s.

Our read

Wealth platforms doing model risk diligence get a dated validation result against published thresholds, including a fail, which shows motif holds back surfaces that miss its gate. Buyers should not treat private markets underwriting as available until a passing result is published.

Persona

Sep 22, 2026

ImpactMedium

VerificationVerified

Persona released React Native SDK v3.0.0, adding builder support for creating one time link and Persona Relay inquiries, redirect URI overrides, shared data field mappings and Persona host selection. The release moves to iOS Inquiry SDK 3.9.0, removes the SDK's client side iOS theming APIs and raises the minimum iOS deployment target to 15.0.

Our read

Banks and fintechs with React Native mobile apps can now start privacy preserving Relay checks and one time link verifications directly in the app, but teams must plan migration work for the removed theming APIs and the iOS 15 floor.

TidalWave

Sep 22, 2026

ImpactMedium

VerificationPartially Verified

Tidalwave added HELOC support to its SOLO AI origination platform as a separate product line rather than a setting on the mortgage flow. SOLO's agentic AI runs a HELOC specific borrower intake, keeps the approved credit line and initial draw as separate figures, calculates draw and repayment period math, validates that the initial draw stays within the approved line, records the correct lien position, and passes HELOC specific terms into the LOS.

Our read

Lenders adding home equity volume while refinance demand is weak can run second lien HELOCs through the same AI intake and LOS handoff instead of forcing them through first lien forms and manual rekeying. The post names no live lender deployments, so buyers should confirm availability and LOS coverage (the tech specs list ICE Encompass and Calyx Path).

Verisk

Sep 22, 2026

ImpactMedium

VerificationPartially Verified

Verisk and Equifax introduced Inflection Homeowners, a credit based insurance score built for property insurers that uses trended credit attributes instead of a point in time credit snapshot. It is aimed at homeowners underwriting, quoting and renewal decisions and extends the existing Inflection Auto model to property lines. Verisk states its own analyses show 2.6X lift between the lowest and highest risk bands.

Our read

Homeowners carriers get a bureau codeveloped rating variable, and Verisk's product page says the model has been filed in most states and approved in a majority of those requiring filings, which shortens the path to rate use. The lift figure is vendor stated with no published method, so carriers still need to validate it on their own book.

Xapien

Sep 22, 2026

ImpactMedium

VerificationVerified

Xapien released a native application on the ServiceNow AI Platform, available in the ServiceNow Store with free trials for ServiceNow customers. The app runs a full, sourced background check on any company or person inside ServiceNow workflows and enriches the out of the box risk scores in ServiceNow Third-Party Risk Management, with configuration for other modules.

Our read

Banks and insurers that run onboarding and third party risk in ServiceNow can pull Xapien adverse media, sanctions and ownership findings into the risk record without a separate tool or manual rekeying. It gives compliance teams an auditable due diligence result attached to the TPRM workflow where the approval decision is made.

Prism Data

Sep 22, 2026

ImpactMedium

VerificationVerified

Prism Data's analytics suite, including CashScore, Income Intelligence and Insights Attributes, is now available inside Engine by Gen's embedded finance platform. Financial product providers on Engine's 1,300 partner network can apply Prism's cash flow analytics to bank data consumers have permissioned on the Engine platform, for credit evaluation, income assessment and product matching without extra applicant steps.

Our read

Lenders already on Engine can use cash flow scores and income data at the marketing and prequalification stage, before a formal application, without building a separate Prism integration.

Psympl

Sep 22, 2026

ImpactMedium

VerificationVerified

Psympl and Experian published results from FreeStar Financial Credit Union, which used Experian Marketing Attributes to select 10,000 prospects and Psympl Motivation Mindsets to tailor a 4 percent high yield savings campaign by direct mail and geotargeted digital ads. FreeStar reported $690,000 in new deposits within two weeks and $1.2 million within 30 days, a stated 3X return. It has since extended Motivation Activation into lending, reporting a 337 percent increase in first mortgage applications and an average loan size increase of more than $132,000, 2025 versus 2026.

Our read

This is a named credit union with dollar and application figures, which lets banks and credit unions weigh psychographic targeting against their own deposit and lending campaign costs. The numbers are self reported by the institution and vendor, with no control group described.

Socure

Sep 22, 2026

ImpactMedium

VerificationVerified

Socure corrected feature logic in the Sigma Identity Fraud and Sigma Synthetic Fraud models that treated last 4 SSN or ITIN inputs as full 9 digit values, which generally produced riskier scores. Score distributions for 4 digit inputs now shift slightly lower and align with 9 digit and missing SSN inputs, with stated gains in precision and recall.

Our read

Lenders and card issuers that collect only the last 4 SSN digits will see score distributions move, so fraud strategy and model risk teams should recheck cutoffs and approval rates for that segment. It also removes a source of excess declines or step ups for applicants who supply a partial SSN.

Bretton AI

Sep 21, 2026

ImpactMedium

VerificationPartially Verified

Bretton published a benchmark of its Agents V2 architecture against Agent V1 on two 250 subject sets (KYC enhanced due diligence with 7 checks, KYB due diligence with 13 checks), with every run graded by its Bretton Evaluator LLM judge on four pass or fail criteria. It reports errors down 37% on KYC and 23% on KYB, perfect score reports up from 42% to 59% and 26% to 38%, and no added run time.

Our read

Model risk teams get a stated test design (same subjects and skills, only the architecture changed, subjects with planted defects and name only plants) they can ask to see, though the grader is Bretton's own LLM judge rather than an independent validator.

Dotfile

Sep 21, 2026

ImpactMedium

VerificationVerified

Dotfile rebuilt Analytics around three read only dashboards: an executive Overview (approval rate, time to decision, automation rate, portfolio risk), AML & Screening (alert resolution times, false positive rate, hit review outcomes) and Documents, ID & eKYC (pass rates and rejection reasons). Users can ask the Autonomy agent in plain language for any other chart, and Autonomy only saves it to a dashboard after the user clicks Approve. Every figure opens the underlying case and check records, which can be exported as CSV or XLSX (up to 500,000 rows) or as a dashboard PDF, with new role permissions for access and custom dashboards.

Our read

Compliance teams can now pull automation rate, false positive rate and time to decision straight from the system of record for auditors and boards, with each number traceable to the cases behind it. The approval step before Autonomy writes anything shows how the agent's actions are kept under human control.

Incode Technologies

Sep 21, 2026

ImpactHigh

VerificationVerified

Incode's platform release, in production from 21 September, adds a GovDataMatch provider for US driver license verification against AAMVA's updated DLDV 3.0 service, with fuzzy name and address matching, AKA names, driving privilege status and indefinite expiration handling, plus two new result fields in the API. As a breaking change in the same release, the capture endpoints now reject uploaded images unless a flow explicitly allows manual ID upload or the new selfie upload from gallery setting, which is off by default and carries a fraud risk warning.

Our read

US lenders and fintechs verifying driver licenses should match more real customers who use name variants, and inconclusive state answers are now kept apart from failures. Blocking gallery uploads by default closes a common route for injected or recycled images.

Jack Henry

Sep 21, 2026

ImpactMedium

VerificationVerified

Jack Henry added two APIs to its Digital Core developer documentation on 21 September. A SilverLake Adapter API in the General Ledger API covers the branch, product, cost center and transaction codes used when posting general ledger transactions to the SilverLake core, and a new transaction disputes API lets developers create, search and fetch dispute records.

Our read

Banks and credit unions moving to Jack Henry's cloud native core can keep general ledger posting in step with SilverLake during the transition. Fintech partners can open and track disputes through the API instead of by hand.

Linkurious

Sep 21, 2026

ImpactMedium

VerificationVerified

Linkurious Enterprise v4.3.13 makes Copilot answers cite the specific nodes and edges they draw on and lets analysts inspect the Cypher query Copilot ran against the graph. The same release locks local user accounts after too many failed logins and can stop users reusing previous passwords. Copilot is still labeled beta and runs on a model the customer supplies.

Our read

Investigators can check a Copilot answer against the graph records behind it before relying on it. The new account controls matter for deployments that do not sign users in through SSO.

nCino

Sep 21, 2026

ImpactMedium

VerificationPartially Verified

nCino published a case study on OSB Group, a UK specialist mortgage lender that rebuilt its lending platform on nCino with Mambu and Publicis Sapient. It reports mortgage offers in under 2 hours from application, AIP decisioning in under 10 minutes, offer to completion cut from 44 to 22 days, and broker fields required for AIP cut from 100+ to fewer than 50.

Our read

Gives lenders named time to decision and time to completion figures from a live UK deployment; the earlier December 2025 account cited offers in under 24 hours, so the 2 hour and 44 to 22 day figures are new. Results are shared with Mambu and Publicis Sapient and no measurement method is stated.

Ondato

Sep 21, 2026

ImpactMedium

VerificationVerified

Ondato released Mobile SDK v3.6.0 for Android and iOS, which its release notes say brings enhanced document capture, clearer user guidance and configurable consent options. The iOS release notes add that the Autocapture automatic document capture model was migrated to a different provider for accuracy and speed, the FaceTec liveness SDK was updated to 9.3.137, full screen document and face capture guideline pop ups were added, and the consent approve button now unlocks only after the user has had a chance to review the terms.

Our read

Banks, neobanks and PSPs embedding Ondato in mobile onboarding get a changed third party document capture model and liveness component, which can shift pass rates and belongs in vendor and model inventories. The consent gating change gives firms better evidence that applicants reviewed terms before approving.

RavenPack

Sep 21, 2026

ImpactMedium

VerificationVerified

The Bigdata.com MCP server added two tools: find_entities, which resolves organizations, people, places and products to entity IDs by name, filters or both, and list_entity_filter_values, which lists the accepted values for a filter such as a person's employer or a place's region. Companies, ETFs, funds and other securities remain under find_securities.

Our read

AI agents connected to Bigdata.com through MCP can now pin a question to a specific nonsecurity entity before retrieving content, which cuts ambiguity in research and monitoring workflows that track people, organizations or places rather than tickers.

Rogo

Sep 21, 2026

ImpactMedium

VerificationVerified

Rogo added Octus as a configurable data source inside Rogo agents, covering Octus expert analysis, private company financials, capital structures, covenant and restructuring data, credit news, debt documents and structured credit data. Agents can use it to screen credits against a strategy, analyze capital structures and covenant packages, monitor portfolios for credit quality changes, and prepare underwriting and investment committee materials. The integration is available to mutual customers with the appropriate Octus licenses.

Our read

Leveraged finance, private credit and restructuring teams can ground agent outputs in licensed sub investment grade data sourced back to Octus rather than loading it by hand, though access depends on holding a separate Octus license.

Sardine

Sep 21, 2026

ImpactMedium

VerificationVerified

Sardine launched Agent Routines, which let customers schedule any agent in Agent Hub, including the Data Analyst Agent and Custom Risk Agents, to run hourly, daily, monthly or on demand and send findings to a notification inbox ranked by severity. Sardine says Routines and an underlying agent harness, which carries memory between runs and governs which tools and data an agent can touch, are live across its AI agents. Basic agent scheduling was previewed in the 14 Sep Agent Hub post; the named product, severity inbox and harness are new in this release.

Our read

Recurring fraud and AML reviews such as card testing scans, EDD cohorts and sanctions rescreening can now run without an analyst starting them, so buyers need to decide which findings route to mandatory human review and which follow ups the agent may apply on its own. Sardine states each scheduled run is logged with cited evidence, which supports exam and model risk documentation.

SESAMm

Sep 21, 2026

ImpactMedium

VerificationVerified

TextReveal API release 3.33.0 adds universe level metrics routes that return case and event counts with intensity score distributions, plus entity counts grouped by highest UNGC classification, all filterable by ISIC4 sector. It also adds an experimental Universes CES Overview route that aggregates Controversy Exposure Scores for a whole universe over time (min, max and average by period), an ISIC4 sector filter on universe case, event and document routes, an entity annual revenue field, and custom column labels for CSV case and event report exports.

Our read

Asset managers and banks that monitor portfolios or loan books through the API can now pull portfolio level controversy, UNGC and sector exposure in one call instead of aggregating entity by entity. The CES overview route is flagged experimental, so its parameters and responses may change before teams build production reporting on it.

Earnix

Sep 17, 2026

ImpactHigh

VerificationVerified

Earnix launched Agent Hub, a curated catalog of more than 25 insurance specific AI agents and apps inside Earnix AIOS, its AI Orchestration System. Earnix says the agents work within pricing and rating, underwriting and customer engagement solutions, drawing context from policy administration systems, data platforms and underwriting workbenches under defined permissions, traceability and human oversight. Fourteen were demonstrated at Earnix Excelerate London; examples include Model Feature Mapper, which connects model features to the appropriate data variables for pricing and actuarial teams, Product Expert Advisor and Premium Explainer.

Our read

This is Earnix moving from recommending a price to acting inside the workflow that sets it. Pricing is a regulated decision in most markets, so the permissions and traceability claims are the ones to test. Ask which agents can change a live rate, and what record each action leaves for a rate filing.

Elliptic

Sep 17, 2026

ImpactHigh

VerificationVerified

Elliptic launched Decode, an AI agent built on its blockchain intelligence platform that answers natural language questions about cryptoasset addresses and their exposure. Every answer carries Thinking Steps showing how the question was interpreted and the query that produced it, and Elliptic states the same question against the same data reaches the same conclusion. Elliptic positions it as the first solution built to the Elliptic Standard, published a week earlier. It is live with government customers, with private sector availability stated for later in the year.

Our read

Crypto compliance teams have to defend every risk call to an examiner, and an answer without its reasoning is hard to defend. The Thinking Steps are the feature to test: check whether they cite the underlying attributed intelligence Elliptic holds, or only narrate the conclusion. Private sector buyers should note this is not generally available yet.

Bottomline

Sep 17, 2026

ImpactHigh

VerificationVerified

Bottomline launched Global Pay Connect, a SaaS platform that brings global and domestic payment connectivity under one integration. It offers API enabled services, third party data integration and AI enhanced insights for real time visibility into payment activity across networks such as Swift and SEPA.

Our read

Banks running separate stacks for Swift, domestic rails and instant payments pay for that fragmentation with every ISO 20022 change. The question is migration: ask what moves across first, and whether existing Bottomline connectivity customers move automatically or buy a new product.

ACI Worldwide

Sep 17, 2026

ImpactMedium

VerificationVerified

ACI Worldwide added a cloud native Financial Messaging capability to its ACI Connetic payments platform, so payment messaging now runs on the same platform as payment processing.

Our read

Messaging is usually the oldest layer in a bank's payments stack and the hardest to replace. This matters most to banks already consolidating onto Connetic; others should ask exactly which networks and message formats it covers before treating it as a replacement for an existing Swift interface.

Abrigo

Sep 17, 2026

ImpactMedium

VerificationPartially Verified

Abrigo added an AI tool to Abrigo Allowance, its CECL solution, that automates the calculation and reporting of expected credit losses. It writes narratives that explain what changed between calculation periods and which factors drove the movement.

Our read

The narrative is the part auditors and boards actually read, and it is usually written by hand at quarter end. An AI written narrative only saves that time if it is traceable, so ask whether each statement links back to the model inputs behind it.

Aptus.AI

Sep 17, 2026

ImpactMedium

VerificationPartially Verified

Aptus.AI launched Next-OS, a platform that turns legal and regulatory texts into structured data to automate compliance work such as checking rules, tracking amendments and mapping regulatory impact. Aptus.AI reports that one large organization cut legal opinion preparation time by 50 to 60 percent across about 110 opinions a month.

Our read

Regulatory change management is where compliance teams lose the most hours to reading. The efficiency figure comes from a single unnamed client and is vendor reported, so treat it as a claim to test on the buyer's own rulebook rather than a benchmark.

Avaloq

Sep 17, 2026

ImpactMedium

VerificationPartially Verified

CA Next Bank became the first Swiss bank to go live on Avaloq's new SaaS mobile banking app, which is built on an architecture that separates the user interface from the core banking back end.

Our read

A first live client is the point where a new banking app stops being a roadmap slide. Separating the front end means features can ship without touching the core; the proof will be CA Next Bank's release cadence over the next two quarters.

Infrrd

Sep 17, 2026

ImpactMedium

VerificationVerified

Infrrd launched IDPForge, a self serve document automation API for developers. It routes each processing stage to a chosen model, supports bring your own key for model providers, includes an embeddable correction interface, and runs a continuous learning loop scoped to each customer's own data.

Our read

Bring your own key lets a lender keep model spend and data processing under its own provider agreements, which simplifies vendor risk reviews. The learning loop scoped to one customer's data is the other detail to get in writing: it means one lender's corrections never train another's extraction.

CUBE

Sep 16, 2026

ImpactMedium

VerificationPartially Verified

CUBE's regulatory intelligence now powers a new Regulatory Horizon Scanning capability inside IBM watsonx.governance. It monitors AI related developments from regulators, legislators and standards bodies and surfaces them in the governance workflow, where teams can map a change to an AI use case and keep a record of how it was assessed.

Our read

Banks and insurers are starting to track AI regulation alongside their model inventories, mostly by hand. This puts the regulatory feed beside the inventory with a trail of each assessment. It lives inside IBM's product, so it only helps firms already running watsonx.governance.

Napier AI

Sep 16, 2026

ImpactMedium

VerificationPartially Verified

Napier AI made its Continuum financial crime compliance platform available on the Microsoft Marketplace, so Microsoft customers can find, buy and deploy it through their existing Microsoft relationship.

Our read

This is a procurement change more than a product one. Firms with a Microsoft Azure consumption commitment can sometimes buy eligible software against it, so ask whether Continuum qualifies; if it does, a compliance tool purchase that normally takes months can move much faster.

SEON

Sep 16, 2026

ImpactMedium

VerificationVerified

SEON expanded its signal intelligence to detect fraud rings and the fake identities now being produced with generative AI.

Our read

Synthetic identities built with AI can pass document checks one at a time and give themselves away as a group. Ring detection is the right place to look; ask SEON which signals were added and how the change affects false positive rates on existing rules.

Marloo

Sep 16, 2026

ImpactMedium

VerificationPartially Verified

Marloo introduced Ask Marloo, which lets financial advisers and support staff query their whole client book. Answers draw on meetings, documents, emails and other records held across the firm's systems.

Our read

The use case is the inherited book: an adviser taking over clients needs their history without reading years of files. Ask what each answer cites, because an adviser relying on a summary of past advice needs to see the record behind it.

V7 Go

Sep 16, 2026

ImpactMedium

VerificationVerified

V7 Go moved chats onto Go Agent V2, which carries a model picker and a reasoning setting so a lookup and a long analysis need not run the same way. GPT-6 Astra, Gemini 3.7 Flash and Gemini 3.8 Flash are selectable and Gemini 2.5 Pro is retired, and the settings tooltip now names the concrete model rather than a generic label. The same release keeps every turn of a run with each tool call stored beside the result it returned, attributes answers by the text they quote, and records an MCP tool invocation as user activity with the sign in route behind it.

Our read

The named model and the complete transcript are the parts a model risk team should care about. Knowing exactly which model produced an extraction is where any validation record starts, and a platform that swaps models behind a generic label makes that impossible. Attributing a tool call back to a person is the other control a second line will ask for.

Personetics

Sep 15, 2026

ImpactHigh

VerificationPartially Verified

Personetics launched Banking Console, which gives relationship managers a daily ranked list of customers who need attention, why, and what to say. It layers Personetics' customer intelligence onto the CRM the bank already uses, including Salesforce and Microsoft Dynamics, and lets the bank set priorities such as growth or retention that shape each banker's plan.

Our read

Personetics has sold this intelligence into digital channels for years; this puts it on the banker's desk for the first time. The bank configuration page is the lever to examine, because whoever sets those priorities is effectively deciding which customers get a call.

Sumsub

Sep 15, 2026

ImpactMedium

VerificationVerified

Sumsub launched Workforce Verification, which brings document checks, biometric liveness and background screening into enterprise identity and HR platforms. It triggers checks at high risk moments across the employee lifecycle, from hiring to help desk account recovery, and works with tools such as Okta and Ashby.

Our read

A password or MFA token proves someone holds a credential, not who they are, and help desk account recovery is where deepfake social engineering is landing. For banks, this moves Sumsub from customer onboarding into protecting the institution's own staff access.

ACA Group

Sep 15, 2026

ImpactMedium

VerificationVerified

ACA Group introduced a compliance data migration capability for firms moving Code of Ethics and personal trading data between compliance platforms. It supports full migration, export and archive, or a hybrid of the two, so historical records survive the move.

Our read

Switching personal trading platforms usually stalls on history, because regulators expect years of records to survive the change. Ask how ACA validates the migrated records, since a gap found during an exam is the exact risk the migration exists to avoid.

AlphaSense

Sep 15, 2026

ImpactMedium

VerificationVerified

AlphaSense added Clinical Trial and Drug Data, bringing US and EU clinical trial registries, drug labels and patent data into its knowledge graph. Users can ask scientific and commercial questions in plain language and trace each answer back to the underlying registry and label records.

Our read

For investors covering biotech and pharma, trial data usually sits in separate databases from market research. Traceability to the registry record is the detail that matters for anyone putting a trial readout into an investment memo.

Bits Technology

Sep 15, 2026

ImpactMedium

VerificationVerified

Bits Technology added power of attorney support for organization stakeholders during onboarding, and expanded risk classification for Finnish companies to cover all industry and legal form codes. The release also improves the speed and precision of PEP and sanctions screening.

Our read

Mandates are a common blocker when onboarding business customers, because the person applying is often acting for someone else. The Finnish coverage matters most to Nordic lenders; others should ask which markets get the same treatment next.

CSI

Sep 15, 2026

ImpactMedium

VerificationVerified

CSI launched Commercial Suite, a set of payment and cash management capabilities that lets community banks serve business customers.

Our read

Community banks lose commercial deposits to larger banks on treasury tools, not on rates. The question is depth: ask which payment types and cash management functions are included, and whether they run on the bank's existing CSI core.

Riskified

Sep 15, 2026

ImpactMedium

VerificationVerified

Riskified brought its Identity Risk Intelligence into Zendesk Industry Agents. Service teams resolving returns and claims can see fraud and abuse signals drawn from Riskified's network of orders, claims and chargebacks.

Our read

Refund and return abuse is usually decided by service agents who never see fraud data. Putting the signal in the ticket is the right place for it; ask what the agent is expected to do with a high risk flag, so legitimate customers are not caught in the net.

Transient.AI

Sep 15, 2026

ImpactMedium

VerificationVerified

Transient.AI launched Caddie.AI RIA Prospecting for Hedge Fund Sales. Sales teams can ask in plain language for RIAs matching criteria such as city, AUM threshold and hedge fund allocation, and get back firms with CRD number, city, AUM and custodians taken from ADV filings, plus a map view and a filterable grid with export. Transient.AI states coverage of 11,986 RIAs, 2,717 of them pure RIAs, that it runs in the customer's own tenant with zero data retention, and that it is available on the Azure and AWS marketplaces.

Our read

Hedge fund sales teams increasingly target RIAs, and prospecting them means stitching together ADV filings by hand. Running in the customer's tenant with no retention is the detail that gets this past a vendor risk review. The data is public filings, so test recency: an ADV is a point in time document and an allocation flag can be a year stale.

Nasdaq Verafin

Sep 15, 2026

ImpactMedium

VerificationPartially Verified

Nasdaq Verafin released an integration with Stablecore that brings digital asset transaction activity into the Verafin platform, so fiat and digital asset data can be monitored together for financial crime. It is in beta with select customers.

Our read

Banks and credit unions adding stablecoin or tokenized deposit services would otherwise need a second monitoring tool for on chain activity. One view across both is the right design; it is still in beta, so ask which assets and chains are covered today.

Sapiens

Sep 14, 2026

ImpactHigh

VerificationVerified

Sapiens launched the Sapiens Autonomous Insurance Platform, SapiensAIP, a SaaS platform that brings agentic tools into its core policy administration systems across underwriting, policy, billing, claims and customer engagement. Its Migration Hub ingests legacy files as they are, proposes data mappings with confidence scores for experts to review, and keeps each transformation reversible.

Our read

Core system migration is where insurance modernization budgets run over, so the Migration Hub is the most consequential piece here. Confidence scored mappings with human review is the right shape. Ask for results from a completed migration, since the first named customer is still evaluating.

Built Technologies

Sep 14, 2026

ImpactHigh

VerificationVerified

Built's September release for CLA+ adds Automatic Draw Submission in the Sponsored Borrower Portal, which separates, labels and matches borrower documents to lender requirements. It also adds deferred rules, approver groups, finer control over funding notice PDFs, and inspection data inside the draw budget so draws can be checked against verified progress.

Our read

Construction draws are paper heavy and slow, and every day a draw waits costs the borrower. Checking a draw against inspected progress inside the budget is a control lenders have been doing by hand; ask how Built handles a mismatch between the inspection and the request.

Addepar

Sep 14, 2026

ImpactMedium

VerificationPartially Verified

Addepar is one of the connectors in Claude for Financial Advisors, which Anthropic launched on 14 September. Advisors can give Claude governed access to Addepar portfolio data and analytics across public and private markets for research, meeting preparation and follow up.

Our read

Addepar holds the consolidated view of complex portfolios, including private holdings most tools cannot see. That makes it the most valuable connector in the launch and the most sensitive, so firms should confirm which Addepar permissions the connector inherits before an adviser points Claude at a family's full balance sheet.

Zocks

Sep 14, 2026

ImpactMedium

VerificationVerified

Zocks brought its client conversation context into Claude for Financial Advisors. Advisers working in Claude can draw on what clients said in past meetings, alongside the portfolio and CRM data from the launch's other connectors.

Our read

Meeting notes are where the personal context lives: goals, family changes, worries. Feeding them to an assistant makes its briefs far better and raises a recordkeeping question, so confirm how the firm's archiving policy applies to what Claude produces from them.

SS&C Technologies

Sep 14, 2026

ImpactMedium

VerificationPartially Verified

SS&C Black Diamond is one of the connectors in Anthropic's Claude for Financial Advisors, bringing portfolio, performance, holdings and rebalancing data into Claude for advisers who use Black Diamond.

Our read

For Black Diamond firms, this puts portfolio reporting data in front of an assistant that can also draft the client letter. Rebalancing is the sensitive part: check whether the connector only reads, or can stage trades for review.

Jump

Sep 14, 2026

ImpactHigh

VerificationPartially Verified

Jump introduced AI assisted, real time account opening with Schwab Advisor Services and Fidelity, with the Fidelity workflow running through Dispatch. During a client meeting, Jump pulls information from the conversation, documents, forms and CRM, flags anything missing, and gets the application reviewed and sent for signature before the meeting ends. On Schwab it can also start asset transfers and set up ACH bank profiles.

Our read

Not in good order submissions are the cost of onboarding, and they usually surface days after the client has left. Catching the gap while the client is still in the room is the real gain; ask how the advisor confirms fields Jump filled from the conversation before anything is signed.

Jumio

Sep 14, 2026

ImpactMedium

VerificationVerified

Jumio extended selfie.DONE, its reusable identity service, to Asia Pacific. Users who have already enrolled can pass later identity checks with a selfie alone, without submitting their documents again.

Our read

Reusable identity cuts friction for returning users, but it also means the first verification carries every later one. APAC firms should ask how often the underlying document check is refreshed, and how an expired or revoked ID gets caught.

Conquest Planning

Sep 14, 2026

ImpactMedium

VerificationPartially Verified

Conquest Planning opened direct access to its Verifiable AI Financial Advice Engine to independent advisers and RIAs, who could previously reach it only through large enterprise institutions.

Our read

Independent RIAs can now license the same planning engine the large firms use, without an enterprise agreement. Ask how pricing works for a small practice, and whether the engine documents its recommendations well enough to support the firm's own best interest obligations.

Kore.ai

Sep 14, 2026

ImpactMedium

VerificationPartially Verified

Kore.ai released Agent Platform v1.6.3, which adds a command line tool for automating the agent lifecycle, support for the Microsoft Foundry model catalog, and PII detection by entity. It also routes all outbound traffic through configurable proxies and adds evaluation suites for each environment.

Our read

Proxy control of every outbound call and entity level PII detection are the items that usually hold up a bank's security review of an agent platform. Evaluation suites per environment let teams run the same checks in staging that they will run in production.

Incode Technologies

Sep 14, 2026

ImpactMedium

VerificationPartially Verified

Incode's platform release, deployed to its Demo environment on 14 September, adds eKYC data sources for the United Kingdom, the Netherlands and Finland. The same release removes a deprecated identity search endpoint, which now returns an error.

Our read

New eKYC sources let Incode check identity against authoritative records in three more European markets, not just against the document. Engineering teams should move off the retired search endpoint now, since a Demo change normally previews production.

Aveni

Sep 11, 2026

ImpactMedium

VerificationPartially Verified

Aveni upgraded Aveni Detect, its compliance monitoring platform for advice and customer interactions, with case level outcome testing and population level trend reporting. Firms can configure assessments against specific priorities, including Consumer Duty, customer vulnerability and complaints handling.

Our read

Consumer Duty shifted UK firms from proving a process was followed to proving an outcome was reached, and case level testing is the evidence format that regime asks for. The population view is what turns individual call scores into something a board can read, which is where most monitoring tools stop short.

CENTRL

Sep 11, 2026

ImpactHigh

VerificationVerified

CENTRL's September release adds approval controlled actions to its CentrlX agentic platform: the agent proposes a field update and a user reviews and approves it from a chat card before anything changes. The release also adds reusable instruction skills, report generation from client templates, structured list answers for assessments, and automated monitoring of SEC Form ADV filings for followed managers.

Our read

Proposal then approval is the pattern a due diligence or vendor risk team can defend to an auditor, because every write has a named approver and a moment of review. The skills and templates are what make that usable at volume; an agent that needs re instruction on every assessment is a chat window, not automation.