Directory of AI investment research platforms for asset managers

Aiera tracks corporate and economic events for asset managers, hedge funds, investment banks and research publishers. It turns earnings calls and other events into live transcripts, summaries and searchable research. An AI research assistant answers questions across broker research, Third Bridge expert insights, transcripts, filings and news, with every source attributed. Firms use it through dashboards and a mobile app, or pull its content into their own tools through enterprise interfaces and a model context protocol server. Its content also appears inside research platforms such as BlueMatrix and in Claude for Financial Services. Aiera states its human reviewed transcripts are 99.9 percent accurate. A consortium of research providers backs the company, including Bank of America, Citi and UBS. Ken Sena is chief executive and cofounder.

AllMind is a research terminal for institutional investment teams at asset managers, hedge funds, sell side research desks and corporate investor relations groups. Its financial ontology resolves every mention of a company, whether in a filing, an earnings call or a firm's own note, onto one object, then links that object to suppliers, estimates, guidance and theses. Behind it sit what AllMind counts as more than 6,800 licensed datasets and 750 million documents. Chat answers questions with passage level citations. Grids runs one question across hundreds of tickers with a cited answer in every cell, and Reports drafts earnings reviews, IC memos, comps analyses, credit notes and initiation style primers in a firm's house template. Agent Studio runs deep dives in the background, schedules recurring briefs and checks a stated thesis against each new filing, transcript and headline. Licensed content comes from FactSet, S&P Capital IQ, LSEG, MSCI, Databento, Aiera, Quartr and BlueMatrix, with ratings from Fitch, Moody's, DBRS and S&P Global Ratings, across 18 markets in North America and Europe. Named clients include RBC Global Asset Management, Sagard, Armistice Capital and Algonquin Capital. AllMind has a SOC 2 Type II report, hosts data in Toronto, Virginia and Iowa, and says customer content never trains its models. AllMind is headquartered in Toronto. Anwaar Malik, its cofounder and chief executive, announced the launch in July 2025.

AlphaSense is a market intelligence and search platform for investment banks, hedge funds, private equity firms, asset managers, consultancies and corporate strategy teams, founded in 2011 by Jack Kokko after his time as a Morgan Stanley analyst. It combines natural language search and generative tools with a large licensed content estate spanning equity and broker research, company filings, event transcripts, expert call transcripts, news, trade journals and a client's own internal research. Generative features include summaries of every earnings call with sources cited, document level question answering, a grid product that runs the same question across many documents, financial data, and workflow agents. The company states it serves more than 6,500 organizations including 88 percent of the S&P 100, passed $500 million of annual recurring revenue, and employs roughly 2,900 people. It reached a $4 billion valuation alongside a $650 million round co led by Viking Global Investors and BDT and MSD Partners with JP Morgan Growth Equity Partners, SoftBank Vision Fund 2, Blue Owl, Alkeon, Alphabet's CapitalG and Goldman Sachs Alternatives participating, and has bought three companies to widen the content and tooling: Sentieo in 2022, Tegus for $930 million in 2024, and Carousel in 2025. It runs a trust center carrying SOC 2 Type II attestation and ISO/IEC 27001:2022 certification with bridge letters and a published policy library. It offers bring your own key and bring your own storage deployment alongside customer managed encryption keys, and mirrors a customer's existing permissions so the assistant cannot surface content a user is not entitled to see. It states that its AI is never trained on customer data and that it works only with model providers enforcing zero data retention. It also states plainly that it is not currently certified to ISO/IEC 42001 while pursuing it and committing to publish progress.

Artian builds multi agent AI for financial institutions. Banks and other financial services firms use it to build, deploy and supervise agents that handle operational work in areas such as capital markets, wealth management and operations. An agent builder lets staff create agents in plain language, workflow agents run structured or dynamic plans, and an execution platform lets agents collaborate through what Artian calls a multi agent exchange. An agent registry gives oversight of deployed agents, and Artian says humans supervise the agents and that deployments follow the customer's software development lifecycle. Artian names no customers, citing a US top 10 bank and a global top 10 bank. It is based in New York.

Auquan builds autonomous agents for the knowledge work that occupies analysts at asset managers, investment banks, private equity and private credit firms and insurers. Combining an agent architecture with retrieval augmented generation, it takes disorganised inputs spread across incompatible formats, document types and languages and produces finished outputs: investment memos, credit prescreens, responses to requests for proposal, business involvement and know your business screening, sustainability performance reporting and limited partner reporting. Its credit agent is presented as executing private credit workflows end to end. The company positions itself against horizontal generative tools aimed at shallow search and support tasks, targeting instead the multistep work that takes an analyst days.

Binocs produces the analysis a deal process needs, using agents organized like an investment team. Private equity, private credit, consulting, investment banking and corporate development teams are the buyers. Outputs include industry reports, investment memos, deal screening, sell side CIMs and credit memos, each backed by citations. Customers can run it self serve or add an optional layer of human experts. Logos on its homepage include Stout, Kaiser Associates and Vivriti. Binocs states SOC 2 compliance and AWS hosting, and says its in house and open source models do not train on client data without consent. It operates from Bengaluru and Jersey City.

Bipsync sells a research management system to institutional investment teams, positioned as the place where an investment firm's accumulated judgment lives rather than as an analytics engine. The platform ingests research automatically from email, documents and third party research providers, and organizes it across teams with optional model assisted tagging. Notes, memos, meeting records, diligence material and decisions sit in one searchable system of record with configurable workflows, permissions and compliance controls. In April 2025 the company launched a named suite of AI features aimed at extracting usable findings from a firm's own proprietary research rather than from public market data, an unusual orientation in this market. A February 2026 integration with Arch pulls private markets documents directly into the workspace, including quarterly reports, financial statements, capital calls and distributions, classified by fund, manager, asset class and effective date. Bipsync was founded in 2012 by a former hedge fund analyst and is headquartered in New York with engineering in Cardiff. The company states that clients represent over $4 trillion in combined assets and include fifteen of the twenty largest United States university endowments and six of the eight Ivy League endowments.

BlueFlame sells AI deal intelligence software to private equity firms, investment banks, private credit lenders, real estate investors, endowments and hedge funds. It pulls a firm's scattered deal information into one place and uses AI to analyze investment committee memos, review data room documents, personalize outreach and run workflows through its agent, Amp. It connects to CRMs such as DealCloud, Salesforce and Affinity, document stores such as SharePoint and Box, and data sources including FactSet, S&P Capital IQ, Preqin and Grata. It is model agnostic, runs on AWS with each client's data in an isolated environment, and reports a SOC 2 Type II certification. BlueFlame is backed by Datasite and works across its virtual data rooms, where more than 4,000 firms use it. Named clients include Siguler Guff, Anchorage Capital Partners and Stenham Asset Management. It has offices in New York, London and Sydney.

Boosted.ai builds Alfa, an AI research system that watches companies, portfolios and market events and writes sourced briefs and memos. It sells to individual investors, analysts, institutional teams, brokerages and developers. Each kind of buyer gets its own tier. Alfa Prime gives institutional teams agents that argue bull and bear cases against a fund's own criteria and return a memo with a conviction score. Alfa Powered lets brokerages embed the product under their own brand, and the Alfa Platform gives developers APIs ready for the Model Context Protocol. Boosted.ai has built AI for investment research since 2017 and states $5 trillion in client assets supported. It holds SOC 2 Type II and ISO 27001, 27701 and 42001. It is based in Toronto.

Brightwave is a New York AI research platform for investment professionals, founded by engineers from artificial intelligence and financial data infrastructure backgrounds. Its systems synthesise analysis across a corpus of hundreds of millions of documents including securities filings, earnings call transcripts, sell side research, breaking news, market data, expert network calls, investment memos and a customer's own internal knowledge base content, then distil the findings into written narratives structured the way an analyst would write them rather than returning a ranked list of passages. Users can also question primary sources directly. Buyers span the buy side, from owner operated registered investment advisers to crossover hedge funds, with a customer base reported at over $120 billion of assets under management, and wealth management firms use it to prepare client facing market commentary. Funding totals $21 million across a seed round in June 2024 led by Decibel Partners with Point72 Ventures, and a $15 million Series A in October 2024 led by Decibel with OMERS Ventures.

Cohesion sells a team of monitoring agents to fundamental equity investors, tracking the companies an analyst covers across earnings calls, investor presentations, regulatory filings, news, podcasts and social feeds, filtering out noise and reporting the developments that bear on an existing investment thesis. The pitch is aimed at the manual half of an analyst's week, the hours spent querying platforms and reading through sources that older research systems do not capture at all, and the company positions the product as an agentic teammate rather than a search tool. It was founded in 2026 in New York by three people whose backgrounds sit either side of the problem, a hedge fund analyst covering software, an engineer from a cloud provider's security organization, and a builder of investor facing artificial intelligence at a large asset manager. It reports being live with more than ten long and short and long only fundamental equity funds managing a combined $10 billion or more, and offers a free trial in which a prospective fund submits a list of tickers ahead of earnings season.

Crunched is an AI analyst that works inside Excel and PowerPoint, built for the people who live in spreadsheets at investment banks, private equity firms, investment firms and advisory practices. It builds valuation models, business cases and leveraged buyout models from scratch, fills in a firm's existing templates, and links data pulled from information memorandums, annual reports, PDFs, images and web sources into those models before turning the results into slides. Its most distinctive job is review. It checks a workbook of any size for broken links, logic errors and finance specific mistakes, such as a missing add back in a cash flow or a mislinked revenue assumption, so an initial review that used to wait for a vice president happens in minutes. Every number stays live linked from source to model to slide, and any change the agent makes can be undone and reapplied. The founders, Philip Borge and Michael Sakowski, spent more than 10,000 hours in Excel in the private equity practice at McKinsey before starting the company in 2025 with Lars Musaeus and Markus Skagemo. Crunched went through Y Combinator and raised a seed round of about $6 million, with First Round, 20VC and Shorooq Partners among its backers. It is headquartered in San Francisco and runs its engineering from Oslo. Rather than selling one generic agent, it tailors workflows to each firm's own methodology, working alongside customer teams. Its own materials name professionals at firms including Pareto, Nordover Capital, BCG, PwC, BDO and CBRE among its users, and its most recent positioning leads with advisory and audit firms alongside finance.

Endex is a New York company building an AI agent that lives inside Microsoft Excel rather than in a separate platform, on the reasoning that Excel remains the working environment of investment banking, private equity and corporate finance. The agent retrieves, synthesises and reasons over financial data drawn from a firm's own internal files, public disclosures including securities filings, and licensed sources, unifying them into one searchable plane. It builds three statement models and discounted cash flow valuations, extracts from documents, reconciles disagreeing sources, cross checks figures to surface discrepancies and flag restatements buried in footnotes, and can deliver output as an Excel model, a document, a slide deck or an email. Two oversight features are central to the product: integrated citations attributing every output to its underlying source, and workflow tracing that maintains labeled cell references so a user can see which cells the agent produced. Buyers are named by category rather than by firm and span investment banks, private equity managers above $100 billion in assets, global multi strategy hedge funds, Fortune 500 finance teams and listed real estate developers. Funding stands at roughly $14 million, with the OpenAI Startup Fund among the backers.

Evalueserve is a global research, analytics and operations firm founded in 2000 and headquartered in Zurich, delivering knowledge work to clients across several industries from delivery centers in India and elsewhere. Financial services is a separately addressed vertical with its own lines covering corporate and commercial banking, investment banking advisory, investment management and research, lending services, and risk and quantitative work. Beyond analyst time, it also ships named, subscribable products. Spreadsmart automates financial spreading for commercial lenders, converting borrower disclosures into structured credit inputs, and was highly commended as a lending technology provider at the 2024 Banking Tech Awards while the firm won best technology provider at Credit Strategy's 2024 Lending Awards. Insightsfirst is a market and competitive intelligence platform combining generative models, machine learning and configurable dashboards with subscription based personalization and user level access control, and it also underpins early warning work that collects and vets large volumes of data to identify borrowers exposed to emerging financial or external risk. A distinguishing feature is Ask an Analyst, which routes a user query from the platform to the firm's own analysts for additional insight or a new study, making the human layer an explicit product component rather than a delivery detail. Peel Hunt is a named client, and its head of research publications is quoted publicly describing an equity research distribution deployment. Independent recognition includes leader placement in Forrester's 2024 wave for market and competitive intelligence platforms, category leadership in Chartis Research's 2024 credit lending operations report, inclusion in the Chartis QuantTech50, and a leader position in an analyst quadrant for generative AI service providers.

Farsight builds AI that produces finished banking deliverables, from pitch decks and confidential information memoranda to investment committee memos, valuation materials and financial models, written in PowerPoint, Excel and Word and kept in sync across all three. Freeform, launched in May 2026, turns a single prompt plus transcripts, financial data, marketing materials and prior deal documents into a complete deliverable, including CIMs of 60 pages or more. Prior Art Match shapes each output to a firm's past work and house voice, and proactive agents draft materials before anyone asks, working from CRM records, past deals and institutional knowledge. The 2025 acquisition of Presentable AI added plain language editing of fully custom slides. Market data flows in from S&P Capital IQ Pro, PitchBook, SEC filings and North Data's European company registers. S&P Global Market Intelligence took a minority stake in July 2026 and plans to sell Farsight's deliverable builder as an add on to Capital IQ Pro later this year. TresVista signed a multiyear agreement in October 2026 to use Farsight across work for more than 400 clients. Each customer runs in its own VPC hosted by Farsight on AWS, in its own cloud or on premises. Samir Dutta, Noah Faro and Kunal Tangri spun Farsight out of MIT. It raised a $16 million Series A led by SignalFire in 2025 and is based in New York.

Finster AI sells a research and agent platform to investment banks, asset managers and private credit teams. Explorer answers open ended analyst questions across filings, research, data and uploaded documents, with citations that open the exact sentence or table cell behind a figure. Tasks run multi step workflows in the background on a schedule or a trigger, such as a new earnings transcript. PowerPoint Automation builds client and committee decks, company primers, strip profiles and CIMs in a firm's own templates, and the Finster Dashboard shows what has changed across each covered company. Finster's engine powers FactSet AI for Banking, which FactSet launched in alpha in March 2026 alongside an investment in Finster. UBS Investment Bank made a strategic investment in August 2026 as part of Finster's Series B. Data comes from FactSet, PitchBook through its Premium Connector, Preqin and MT Newswires, on top of a proprietary pipeline of more than a million documents on over 8,000 companies with deep EMEA and APAC coverage. Finster offers single tenant and containerized private cloud deployments in US and European regions, supports bring your own language model, commits in its terms never to train on customer data and states SOC 2 Type II compliance. Sid Jayakumar founded Finster in 2023 after seven years of AI research at Google DeepMind. FinTech Collective led its Series A in 2025, and the company is headquartered in London and expanding in New York.

FinTech Studios runs an AI intelligence engine that turns news, filings, regulatory updates and market data into cited research for financial professionals. Compliance teams, analysts, developers and AI agents use it through daily briefs, search and its RegLens regulatory product. Every insight carries an inline citation with the publisher, date and link. Users reach it through the FTS Studio workspace, APIs and SDKs, an MCP server for tools such as Claude and ChatGPT, and connections to Salesforce, HubSpot and Slack. A free plan needs only a login, and Pro starts at $20 a month. FinTech Studios lists S&P Global, Citi, J.P. Morgan, PwC and MSCI among its clients and states SOC 2 Type II. FinTech Studios, Inc. is based in New York.

Hebbia analyzes large sets of private documents for asset managers, private equity firms, investment banks, credit investors and law firms. Deal teams use it to ask the same questions of hundreds of documents at once. Its Matrix product lays documents out as rows and questions as columns, and agents fill in every cell by reading each page. Every answer cites the exact page, paragraph and sentence it came from, and analysts can flag or overwrite any cell. Sources range from SEC filings to FactSet, PitchBook and a firm's own SharePoint, and OpenAI's customer story on Hebbia names the OpenAI models it uses. Customers include Morgan Stanley, Centerview Partners and MetLife. Hebbia lists SOC 2 Type 2, ISO 27001 and ISO/IEC 42001, says neither it nor its model providers train on customer data, and offers United States or European Union processing with dedicated tenants.

Hudson Labs, founded in 2019 in Toronto and formerly called Bedrock AI, builds finance specific language models for institutional equity research. Its founders are Kris Bennatti, a chartered accountant and financial data scientist published by the Harvard Law School Forum on Corporate Governance, and Suhas Pai, who wrote a book on designing large language model applications. The original product was systematic forensic risk analysis, reading the unstructured text of regulatory filings to surface accounting and governance concerns that standard screens miss, and in September 2025 the company launched the Co-Analyst, a research platform producing source verified summaries, guidance identification, multi document comparison and auto generated investment memos. The company frames its pitch against general purpose assistants, arguing they fail precisely where institutional investors need reliability most, on multi document analysis, guidance identification and numeric precision, and describes its own approach as proprietary retrieval, custom models, source selection, noise suppression and in house pipelines. It states the launch followed beta testing with hedge funds, asset managers and family offices representing over $1 trillion of assets under management, with early results showing research time cut by more than half, and it serves asset managers, insurers and securities law firms. A high yield bond analyzer is in development. Investors include Y Combinator, FoundersX Ventures, Born Capital, Zillionize, Brighter Capital and Immad Akhund, and named partners include Wolfpack Research and Enhancing Capital.

Linvest21 builds AlphaCopilot, an investment platform run by specialized agents including equity analysts, portfolio managers and virtual chief investment officers, sold to investment institutions and aimed at giving medium and smaller asset managers capability previously affordable only to the largest firms. It covers the chain end to end, from idea generation and research through portfolio construction, risk management and performance analysis to customized client communications, and a 2025 addition delivers real-time global macroeconomic analysis in fifteen structured sections for chief investment officers and research leads. The company describes a human in the loop architecture in which the models recommend and people decide, and reports hundreds of billions of dollars in assets under its platform's supervision. Its founder spent 22 years at a major asset manager, latterly as chief technology officer of its global investment platform, and holds six financial technology and applied AI patents.

Model ML builds an AI workspace and finance agent for investment banks, private equity firms, asset managers, and large accounting and consulting firms. It produces client ready documents, spreadsheets and decks in each firm's house format. The agent works inside Excel, PowerPoint and Outlook and can run on a schedule or when an event occurs. It reads email, filings, CRM records and data from providers such as S&P Capital IQ and FactSet, using models from OpenAI, Anthropic and Google. Named users include HSBC, Santander and Deloitte. Model ML offers self hosted, single tenant and hybrid deployment, states ISO 27001 and SOC 2 Type II, and commits never to train its models on customer data. It has offices in New York, San Francisco, London and Hong Kong.

Needl.ai is a private enterprise AI platform for financial markets, founded in 2019 by Vikram Srinivasan and Kuntal Shah, with engineering in Bengaluru, India and a US entity, Needl Inc., registered in San Francisco. It positions itself as a context layer rather than a model builder, running inside a customer's own environment and answering questions from that firm's own material: email, drives, wikis, virtual data rooms, CRM records, subscribed research and internal notes, alongside filings, news and roughly 20,000 monitored external sources. Four product capabilities are sold: knowledge management with cited answers, report automation covering filing summaries and credit memos, market intelligence with ranked and routed alerts, and a causal intelligence layer that models the drivers behind market moves rather than scoring them. Eight named workflows sit on top, among them automated due diligence with a cited investment committee memo, credit rating report generation, compliance event monitoring across regulators, a relationship manager copilot checked against mandate and suitability, financial modeling populated from filings, analyst note generation, real time trading intelligence and portfolio monitoring. Buyers span private equity and private credit, asset management, banking and capital markets, wealth management and private banking, and credit rating agencies. Named customers include GrayArch Partners, Scotia Wealth Management, Oaklane Capital Management and MSquared, with unnamed deployments described at a large consultancy, a major ratings agency, a credit agency running $750 million of assets and a trading desk running $10 billion. The company reports roughly 45 staff and holds seven patents on its retrieval and reasoning engine. Its disclosure is detailed for a firm of its size. It names five foundation model families as interchangeable commodity infrastructure and supports customers bringing their own model. It names eleven source systems it connects to and enumerates ISO 27001, SOC 2 Type II and CASA with a Google CASA assessment. It states that it never trains shared models on customer data, and it cites a placement on the S and P AI Benchmarks Long-document QA leaderboard run by Kensho. Set against that, its marketing carries a headline claim of no hallucination and accuracy figures quoted as high as 100 percent and above, which is not a figure that can exist.

Obin AI builds what it calls an agentic workforce for financial institutions, deploying agents that run defined workflows end to end rather than producing drafts for a person to finish. The platform targets continuous monitoring, underwriting at scale and earlier risk detection across private credit, equity, lending and insurance, and its central design claim is that agents operate inside a firm's own controls and audit boundaries, encode that institution's specific logic and decades of accumulated context, and produce traceable and inspectable outputs. The architecture is described as open and free of lock in, with the enterprise retaining full ownership of the intellectual property. Founded by a former head of artificial intelligence at a major global bank and a former Google executive, it emerged from stealth in March 2026.

Primer is an AI analyst for equity investors, built by former analysts to take on the research work behind a buy side investment view. It runs in the browser with research sources already included, from filings, transcripts and consensus estimates to market and alternative data, so a team does not have to wire up each source itself. It reads the source material, pulls out the drivers, builds the financial model, writes the research and can monitor a coverage list on a schedule, flagging what changed without waiting to be asked. Rather than working cell by cell inside Excel, it builds models on its own modeling infrastructure and converts the finished model into Excel, and it learns each team's coverage, preferences and standards over time. The company is Kernel AI Ltd, trading as Primer and registered in London. Its team includes founder and portfolio manager Seb Jory, cofounder Ruggero Gargiulo, a former analyst and quant, head of applied AI Alistair Smallwood, and chief technology officer Chris Mingard. Primer publishes its evaluations and names the model underneath, GPT-5.5, which it says scores 44.3 percent on the BigFinanceBench leaderboard on its own and 79.1 percent inside Primer on the benchmark's public questions. It also reports a backtest of its earnings previews in which trading on its strongest calls produced 34 percent alpha against the S&P 500 over 203 trading days.

Quartr structures the first party material public companies publish about themselves, live earnings calls, real time transcripts, conference and capital markets day audio, filings and slide presentations, and sells it in two forms: Quartr Pro, a research platform where investment professionals query across company materials with every answer linked back to the original source, and Quartr API, a data feed that other research platforms and artificial intelligence systems build on. Coverage runs to more than 13,000 public companies across over 65 markets, with a single interface carrying consistent identifiers across audio, transcripts, filings and slides, and entity structuring across companies, events, people, products, topics and reported metrics linked over time. The company describes itself as infrastructure for company research rather than as an analysis product, and its stated argument is that investor relations material was the primary input to qualitative research and had never been made queryable at scale. Headquartered in Stockholm with offices in New York and Dublin and led by co founder and chief executive Oscar Kuntzel, it states that more than 700 financial institutions and technology companies use its products, including hedge funds, asset managers, equity researchers and investor relations teams, and it also runs a free consumer mobile app carrying the same event material. It reports triple digit growth and net revenue retention of roughly 120 percent. Funding has come in successive rounds from Altos Ventures, which led an $18 million round in 2026 and is now the largest shareholder, alongside the Nordic bank SEB, Ohman and Yanno Capital.

Rogo is an AI research and analysis platform for investment banks, private equity firms and investors. Analysts ask questions in plain language and get cited answers drawn from filings, transcripts, market and deal data, and the firm's own documents. Its agent, Felix, carries deal work from start to finish, producing Excel models, investment memos, diligence materials and slide decks. Data comes from providers such as S&P Global, FactSet and Crunchbase, and the platform runs on OpenAI models. Rogo states more than 50,000 bankers and investors at more than 350 institutions, and quotes executives at Truist Securities, Nomura and Baird. Its security page says customer data sits in siloed environments and never trains its models, and lists SOC 2 and ISO 27001. Rogo Technologies, Inc. is based in New York.

Rowspace builds AI for credit, private equity and other investment firms that learns a firm's own history and how it decides. It maps the firm's documents, models, research, accounting, trade and cash data so that judgment can be applied to new decisions. It runs inside the customer's own cloud, with the customer's keys and access controls, and delivers answers through chat, dashboards and tools such as Excel. Its pages describe clients by type, such as a credit originator monitoring CLO compliance and a private equity firm drawing on 50 years of deal history. Rowspace states SOC 2 Type II and that customer data never trains anyone else's model. It is based in San Francisco.

Samaya AI builds research agents for investment banks, sales and trading desks, hedge funds and asset managers. Its pitch is that in this work, being right matters more than sounding fluent. The agents monitor filings, broker research and social media, write sector reports, build presentations from a firm's own documents and answer complex questions with cited evidence. They run on Samaya's own finance language models combined with frontier models. Morgan Stanley is a named customer across research, sales and trading, and banking, and Samaya states more than 10,000 seats. It states SOC 2 compliance and offers virtual private cloud and on premise compatible deployment. Its founders came from Google Brain and Stanford.

Shortcut is an AI agent for Microsoft Excel built for finance professionals, made by Fundamental Research Labs, a San Francisco applied AI company formerly known as Altera. It runs as an Excel add in, on the web, in a Windows desktop app, through a command line tool and through an API, with a Google Sheets plugin included. An analyst describes the work in plain language and the agent plans it, builds it and checks it, producing models made of live formulas rather than pasted values, so every number can be traced to the logic behind it. Typical jobs are three statement models, discounted cash flow valuations, leveraged buyout models with debt schedules and equity waterfalls, and the refresh and cleanup work that fills an analyst's week. Its users are investment bankers, private equity and hedge fund teams, commercial real estate analysts and accounting teams. TresVista, which supports more than 350 clients across private equity, credit, real assets, public markets and banking, embedded the agent in its modeling and research work after a 45 day pilot. Two independent tests of AI modeling tools published in 2026, by Wall Street Prep and by the Office of the CFO practice at Ankura, each ranked Shortcut first against Claude, Microsoft Copilot and ChatGPT working inside Excel. Both also found that no tool yet produces a model ready to use without a person reviewing it. The parent company raised a $33 million Series A led by Prosus in 2025, bringing its total funding above $40 million.

Terminal X, built by Project Pluto Inc, is an agentic research platform for institutional investors, founded in New York in August 2022 by chief executive Hyun Hong, who spent a decade in mergers and acquisitions, private equity and portfolio management, and chief technology officer Kibum Kim, previously more than ten years at Google as a tech lead across Chrome, Search and the TensorFlow team. Its argument is that general purpose assistants fail investment work because they know nothing about a firm's own thesis, risk limits, investment committee preferences, trusted data vendors or memo conventions. The platform therefore indexes a firm's proprietary material alongside public sources, parsing Excel models at cell level, emails, past deal memos and subscribed research into sentence level retrieval, and combining that with more than a hundred million external sources including filings, earnings transcripts, investor presentations, financial supplements, broker research and news. It supports memo writing, due diligence, thesis validation and portfolio monitoring, and separately publishes data interfaces for market intelligence and event narratives. Its architecture is deliberately model agnostic: a query is decomposed into hundreds of discrete micro steps, each evaluated in milliseconds for the nature of the task, the reasoning required and the quality against latency tradeoff, then routed to whichever frontier model is performing best for that input across the GPT, Claude and Gemini families, with new models added as they launch and prompts hand tuned by former bankers and hedge fund managers. Buyers span hedge funds, private equity and buyout funds, asset managers, family offices, venture firms, sovereign wealth funds, pension funds, securities firms and corporate finance teams across the United States, Japan and Korea. It raised a Series A including DG Daiwa Ventures, the joint venture between Digital Garage and Daiwa Securities Group.

Tracelight is an AI platform for building, checking and presenting financial models, used by deal, valuation and diligence teams at private equity firms, investment banks, corporate development groups and advisory firms. It works inside Excel, PowerPoint and Word through add ins, and on its own web platform. An analyst can ask it to build an operating model, a leveraged buyout or a discounted cash flow valuation to the firm's own template, refresh a comparable companies table, map a trial balance into financial statements, or turn a live model into an investment committee deck in which every figure is cited to its source. Its model review feature audits a workbook for broken links, wrong references, sign flips and other logic faults, and every edit the agent makes arrives as a changeset a person can inspect. The company was founded in London in 2024 by Peter Fuller, Aleksander Misztal and Janek Zimoch, and raised a $3.6 million seed round led by Chalfen Ventures in 2025. Customers shown on its site include Arma Partners, Puma Growth Partners, Vistra, RSM and Armanino. It publishes its own benchmark, run across 80 real Excel models, comparing its agent with the leading general models from Anthropic and OpenAI on building formulas and finding planted errors. It sells on security as much as on speed: a SOC 2 Type II audit, a contractual promise never to train on customer data, and a choice of hosting region for enterprise customers.