Model ML
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.
Capability Axes
Capability grades
15 of 15 axes rated · 10 graded A or B
Agentic modules run continuously or on event triggers, reading email, filings and third party datasets to draft deliverables such as a first investment committee memo from an email thread. Only connectors into data the firm already held would be left behind.
Modules run on a schedule or an event with no person initiating them, landing output as editable documents where a banker works. Approval gates, confidence thresholds and review before a deliverable leaves the firm are not described.
Web intelligence from Perplexity Sonar arrives with citations, confirmed by Perplexity's own case study, and a document review function sits in the product. Output in a firm's prior house formats makes machine work look like a person's, and accuracy figures, evaluation method and model documentation are absent.
A global rollout at HSBC and firm wide rollouts at FT Partners and PwC are reported, and a $75 million Series A led by FT Partners closed in November 2025. No customer is joined to a measured outcome such as hours saved or error rates.
Customer data never trains the models, per the security page, and data can sit in an environment fully isolated from the vendor and its other customers. Retention at the model providers and the wider data lifecycle are not stated.
Confidential deal material, email and documents can stay within infrastructure the firm controls, and GDPR compliance is stated for the product. The security and reference pages show no retention schedule or subprocessor list.
ISO 27001, SOC 2 Type II and AES 256 encryption at rest are named, with a trust center at trust.modelml.com. Audit periods and certifying bodies are not given on those pages.
Supervisory instruments are not named as design targets, though the advisory board carries deep banking experience and no license is needed. How generated pitch materials and memos enter supervised records, and how competing clients on one transaction stay separated beyond siloed storage, are not addressed.
Companies and transactions are the subjects, so the question is accuracy governance, and a draft memo assembled from email threads and private company databases feeds a decision to commit capital. Accuracy reporting, known failure modes and the check between an autonomous module and a client document are unpublished.
Deliverables land as drafts a professional reviews before anything reaches a client. A wrong figure carried into a committee decision by an autonomous memo meets no accuracy guarantee, remediation term or published error rate.
OpenAI, Anthropic and Google models are named down to version on the reference page, with Perplexity Sonar for web intelligence. Exposure of confidential deal content to those providers under the hosted option is left unaddressed.
S&P Capital IQ, FactSet and PitchBook feed in, and output lands inside Excel, PowerPoint and Outlook as documents in the firm's exact prior formats. Which customer relationship and deal systems connect in production is not listed.
Self hosted, single tenant and hybrid deployment are offered, including single tenant on Microsoft Azure and deployment to the client's own infrastructure. Region lists for the hosted option, and where calls to model providers are processed, are not published.
Firm wide enterprise deployments carry no published rate, tier or seat cost. Whether pricing follows seats, modules or output, and whether data connections need the firm's own license, is not indicated.
Investment banks, private equity and asset managers sit beside accounting and advisory firms such as Deloitte, Moelis and BDO, with offices in New York, London and Hong Kong. Lending, deposits, insurance and retail financial services fall outside the deal and advisory workflow covered.
Product updates
Material product, regulatory, evidence and commercial changes at Model ML, each verified against a live source and tagged to the capability axis it bears on. Funding rounds and awards are not product changes and are not logged.
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.
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.
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.
Compared With
Most editorial comparisons pair two vendors the index assesses as direct competitors for the same buyer. Some pair vendors that are adjacent rather than rival, where the useful question is where one ends and the other begins. Each carries a verdict, the buyer conditions that favor each vendor, and a graded side by side.
Alternatives to Model ML
Vendors in the same categories as Model ML whose profiles look most alike, each with a line on what the product does. Similar is not the same: check that each one does the job you need. No vendor pays for placement.
Farsight builds AI that produces finished banking deliverables, from pitch decks and confidential information memoranda to investment committee memos, valuation materials and financial…
BlueFlame sells AI deal intelligence software to private equity firms, investment banks, private credit lenders, real estate investors, endowments and hedge funds.
Finster AI sells a research and agent platform to investment banks, asset managers and private credit teams.
Rogo is an AI research and analysis platform for investment banks, private equity firms and investors.
Rowspace builds AI for credit, private equity and other investment firms that learns a firm's own history and how it decides.
V7 Go is an agentic document workflow platform from V7 Labs, sold into document heavy industries with private markets, insurance and finance among its named verticals.
The index publishes no overall score or ranking. See what we assess and the methodology.
Pricing
Vendor-published figures are labeled as such. Figures labeled “Estimated” are derived from third-party sources and have not been confirmed by the vendor.
No pricing data has been verified for this vendor. Pricing information will be published here once confirmed through vendor disclosure or third-party estimation.