Teslar Software

The platform is workflow, rules and connected core data: exceptions and covenant tracking, pipeline, annual loan review, special assets and portfolio tools that ran for years before any model was added, and a consumer origination system that moves qualifying loans from application to booked in the core in under 15 minutes on automated workflows and decisioning.

AI sits on top in two places, the TALOS assistant and Document Analyzer, under a stated rule of automation where work is repeatable, AI where it needs understanding and people where it needs judgment. The company likens AI to the internet, worth nothing when bought for its own sake, and names quality control, document processing, knowledge, workflows and agent assisted actions as the places it is exploring next.

Each AI feature has a stated stopping point. TALOS works only within the access and officer scoping already set for each user and prepares actions such as exception notices for review; before it changes a record or sends information outside the institution, the user reviews the action, can edit it and must approve it, and its questions, results and actions are logged.

Document Analyzer runs a six step path from arrival through identifying the document, extracting its fields, matching it to the customer, account, exception or checklist item it belongs to and scoring its confidence, so staff review only the files that need it, which Teslar describes as moving document quality control from review by default to review by exception.

Document Analyzer shows a confidence score on each extracted field and a verification step on each match before a document is filed, and TALOS returns its answers with links to the source records in Teslar, so staff can check what the AI found before relying on it. No accuracy figure, validation method or error rate is published for either feature. Documents the system is less sure of are left for a person to review rather than filed automatically.

Liberty National Bank saves an estimated 120 hours a week in the departments using the Exceptions Tracking module alone, and Teslar reports clients clearing exceptions 80 percent faster with its exceptions tools and imaging integration, saving an average of 92 hours a week through workflow automation and running efficiency ratios 14 percent better than peer banks.

Art Stauffer, chief executive of First Texoma National Bank, expects to cut time spent handling loan documents by about 70 percent, freeing almost one full time employee, and Pendleton Community Bank describes a tax return moving from the portal through credit administration into storage without anyone touching it. Bill Loving of Pendleton, Mel Martin of First National Bank of Oklahoma, Pam Lawson of Grand Savings Bank and Chris Floyd of Dream First Bank speak for it by name. The measured outcomes come from the workflow and exceptions tools rather than from the AI features.

The AI boundary Teslar describes is about access: TALOS works within the permissions and officer scoping each bank already sets, so a user sees through the assistant only what that user could already open, and Document Analyzer works on the bank's own incoming documents, active exceptions and checklist items. Whether bank data trains any model, how one bank's data is kept apart from another's and what an outside AI provider retains are not stated. Teslar says its AI capabilities are developed with the security and privacy of sensitive banking information in mind, and that it applies AI only where it meaningfully improves the work.

Teslar's privacy policy, last updated in December 2023, covers its websites and the people who contact the company and expressly excludes the personal information it processes on behalf of client banks as a service provider, so how borrower and account data is handled inside the platform, how long it is kept and who else processes it are not set out publicly.

The platform side rests on security commitments instead: access only for the bank's authorized users on devices the bank approves through multifactor authentication, data protected in transit and at rest, and a System and Organization Controls (SOC) 2 Type II report whose scope includes the confidentiality and privacy of the information processed. Borrowers send documents through an upload box on the bank's own website, and every file is scanned for viruses on arrival.

Teslar offers SOC 2 Type II reporting under Statement on Standards for Attestation Engagements (SSAE) 18 through a partner it describes as one of the most trusted in the industry, without naming the auditor; the report covers the security, availability and processing integrity of its systems and the confidentiality and privacy of the data they handle. Vendor due diligence and vendor management requests go to dedicated email addresses rather than a trust portal.

Controls follow the Cybersecurity and Infrastructure Security Agency's Secure Cloud Business Applications guidance and include intrusion detection and prevention, security event monitoring, a defense in depth design, multifactor authentication from approved devices and end to end encryption, and the leadership team includes a vice president of security and architecture and a vice president of quality and data security.

As a software provider to banks, Teslar states no license or examination of its own; its regulatory weight shows in the work it carries. The special assets system reviews core data, related debt included, to flag loans that should be criticized or classified, gives officers a narrative template designed to meet regulator expectations and can trigger and host work under Accounting Standards Codification (ASC) 310, while deposit tracking covers know your customer and anti money laundering documentation. Its closest brush with a federal program is ppp.bank, which processed more than 20 percent of US Paycheck Protection Program loans the year after its launch.

The stated AI approach governs access and action rather than outcomes: AI works within the access each user already has, consequential actions need user approval and AI activity is recorded. No bias evaluation or named AI governance process is published.

Credit decisions sit outside the AI features: the consumer origination system automates decisioning inside policy driven workflows whose guardrails Teslar says support Regulation B, Regulation Z and fair lending requirements, while TALOS and Document Analyzer find information and file documents.

Responsibility for AI output stays with the bank's own staff by design: nothing TALOS prepares takes effect without a user's approval, the user can edit it first, and its questions, results and actions are recorded. No accuracy commitment or indemnity attaches to the AI, and the one written guarantee Teslar offers, money back if it cannot integrate with a core or imaging system, covers connectivity rather than results. A Grand Savings Bank executive notes that everything done in Teslar carries a time stamp.

The platform runs from Teslar's own Tier 4 data centers, so the hosting layer is known, but neither TALOS nor Document Analyzer names the model or provider behind it, and no subprocessor list is public. Credit bureau, pricing and document preparation connections are listed by type rather than by supplier.

Teslar guarantees integration with any core and imaging system, with the bank's money back if it cannot deliver, and lists integrations by type for core, imaging, email, Microsoft Teams, credit bureaus, pricing and document preparation systems; its most common core and imaging partners appear only as logos, with a note that the providers do not officially endorse it, so none is named in text.

TeslarSync reads from and writes back to the core and imaging systems in real time, handling account and customer updates, payments, stop payments, deposit holds, collateral changes and loan disbursements, and the special assets module unifies data from multiple cores. The loan pipeline connects to document preparation and pricing systems and forecasts funding at 30, 60 and 90 days.

Teslar runs as software as a service in the browser, served from its own data centers in geographically separated facilities that meet Tier 4 requirements, with redundant power, network equipment, climate control and fiber entrances, round the clock monitoring, fire suppression and redundant tier one backbones. Beyond those facilities no cloud provider, region or residency commitment is named.

Borrower facing pieces run inside the bank's own website, the Teslar Unite portal for applications, renewals and documents and the lighter Unite Lite upload box, and the construction module adds Android and iOS apps that lenders and outside inspectors use to send geotagged photos.

Buying runs through a demo request, and no price, tier or unit of charge is published for any module. The catalog is modular, so a bank chooses which lending, credit risk and deposit solutions to take rather than buying a suite.

Implementation follows a four stage method Teslar calls Stellar Start: a systems check of the bank's needs and goals, a launch on a preliminary configuration, optimization that fits the automations to the bank's workflows, and an ongoing phase it calls Orbit in which Teslar keeps working with the bank's team.

Community banks are the whole market, served with consumer lending, commercial lending, credit risk management, portfolio management and deposit operations offers, and with views built for loan officers, loan operations, deposit operations, chief credit officers, chief lending officers and chief executives.

Community Connect carries a bank's consumer lending to the point of sale, letting local businesses such as medical providers, heating and cooling contractors, jewelers and farm equipment and powersports dealers offer financing through the bank rather than a national lender. Named clients run from small community banks to Simmons Bank, the construction module covers loans on one to four family homes, and the business is US only.