Model ML has raised $250 million to build AI that takes over the repetitive analytical work of investment banking, TechCrunch reports.

The pitch is among the least glamorous and most defensible in enterprise software: the grunt work. Pitch books, comparable-company analyses, market updates and the reformatting that surrounds them are heavily rule-based, costly when handled by junior bankers, and repeatable enough that a model can be pointed straight at them.

A quarter-billion-dollar round is a large number for a company selling into a single industry, and it suggests investors believe the market for automating financial analysis is deep enough to sustain it. Banks pay for software that saves analyst hours because those hours are already billed at a premium.

It also continues a pattern that runs through this year's AI funding: capital is moving away from general-purpose chat products and toward tools bolted onto a specific, expensive workflow where the buyer can count the hours saved. Finance is attractive for much the same reason legal and clinical work is — the underlying tasks are high-value, text-heavy and repetitive.

The open question, as with every automation claim in a regulated industry, is how much of the work is genuinely mechanical and how much rests on judgment, client relationships and liability that a model cannot carry. Banks will buy the first part.