Model ML raises $250m at $1.5bn to put AI agents on bankers' desks
General Atlantic and QED co-led the Series B, a bet that the pitchbook-and-spreadsheet grind of dealmaking is the next white-collar workflow to fall to software.

London-based Model ML, which builds AI agents for bankers, consultants and investors, has raised $250m in a Series B at a $1.5bn valuation, co-led by General Atlantic and QED Investors. The round prices the deal-workflow layer of finance as a standalone software market, forcing banks, consultancies and their boards to decide whether to build, buy or wait.
Model ML has raised $250m in a Series B round at a $1.5bn valuation, the London startup said on Tuesday, in a straight bet that the most repetitive work in finance is about to be handed to software. General Atlantic and QED Investors co-led the round. The company builds AI agents for bankers, consultants and investors, a customer set that has historically paid enormous sums for human hours and has been slow, and often right, to trust anything that touches a live deal.
The math is worth pausing on. A $250m raise against a $1.5bn post-money valuation means new investors bought roughly a sixth of the company, with existing holders diluted by a similar amount. That is a large cheque for a Series B. Rounds at that stage have historically clustered in the tens of millions, so the size alone signals that the buyers believe this is a market, not a tooling niche. The cap table is a tell too: Evolution Equity Partners, Airtree, Forestay Capital, Y Combinator, FT Partners, Latitude, Inception Fund and 13Books Capital all participated alongside the co-leads, a mix of growth capital, early-stage venture and financial-services specialists.
Strip the jargon and an AI agent is software that does not just answer a question but carries out a multi-step task on its own: it opens the file, pulls the numbers, drafts the page, checks the arithmetic and hands back something a human can edit. For a banker, that typically means the work that eats a junior's week: comparable company analysis, precedent transactions, data-room summaries, first drafts of a pitchbook or an investment memo. For a consultant, it is research, benchmarking and slide assembly. For an investor, it is screening and diligence triage. None of that is the glamorous part of the job. All of it is billable.
Which is exactly why incumbents should care. Professional services firms run on leverage: a wide pyramid of juniors doing the volume work, a thinner layer of seniors reviewing it, and a partner selling the result. The model works because juniors are cheap relative to what clients pay for their hours. If agents compress the bottom of that pyramid, the arithmetic of the whole firm changes, from utilisation targets to the ratio of analysts to associates to the price of an engagement. That is a board-level question, not an IT question, and it is the reason capital is flowing into this category rather than into another dashboard.
The risks are not small. Banks and advisory firms sit on material non-public information, and client confidentiality is not a nice-to-have; it is the licence to operate. Deployment inside those walls typically means private cloud or on-premise infrastructure, tight audit trails and a clear answer to who is liable when a model gets a number wrong. A hallucinated figure in an internal draft is embarrassing. The same figure in a fairness opinion is a lawsuit. Any vendor selling into this market has to solve accuracy and data residency before it solves speed.
Regulators have been circling. Supervisors on both sides of the Atlantic have spent the past few years signalling that they want to understand how AI is used in advice, trading and risk, and the direction of travel has generally been disclosure and accountability rather than prohibition. That is a workable backdrop for a vendor, but it raises the bar on documentation. Firms that cannot explain how a model reached an output will struggle to put it in front of a client or a regulator, no matter how good the output looks.
London is a logical home for this. The city is one of the few places where a startup can hire people who have actually built a discounted cash flow model for a live deal and engineers who can ship a production system, often within a few Tube stops of each other. It also sits between the US and Asia, which matters when your buyers are global banks. The UK has spent years talking up its ambition to be an AI hub; a $1.5bn valuation on a homegrown, enterprise-focused company is the kind of proof point policymakers like to cite.
For peers, the question is not whether agents arrive in finance. It is who owns the layer. Banks and consultancies can build in-house, buy a point solution, or wait and let a platform swallow the workflow. Each has a cost. Building means competing for the same scarce engineers and carrying the maintenance bill forever. Waiting means ceding the interface, and the data that flows through it, to a vendor that may later sell to your competitor. Model ML now has $250m and a cap table full of people who will expect it to answer that question fast.
This story's Key Insights and Take-aways are locked.
Create a free account to unlock Executive Actions for one credit.
Register to UnlockAlways free for Executives Club members. Join the Club
More in Business
S&P 500 Enters Correction, Down 10% From Its High on AI Worries
The index's slide past the 10 percent line is a symbolic verdict on investor mood, and the AI trade that powered the rally is now the thing testing it.
Dow and S&P 500 enter correction as Nasdaq logs worst week since 2020
All 11 S&P 500 sectors fell more than 1% in the same week for the first time since April 2025, a breadth signal that matters more than any single index level.
Chip stocks lose $2.3 trillion in a week, worst run in 25 years
A $2.3 trillion weekly wipeout in semiconductors is forcing every board with an AI story to ask whether it owns a thesis or just a crowded trade.



