Eve raises to bring AI agents to law firms, targeting the $2T legal market

Summary

Eve, an AI platform built specifically for plaintiff law firms, has raised $103 million at a $1 billion valuation. The company's founder, Jay, says Eve has grown from roughly 13 customers at the start of last year to approaching 500 today.

Why plaintiff, not defense

The company made a deliberate choice to focus on the plaintiff side of the legal market, and the business model logic is straightforward. Defense firms bill by the hour, which creates a perverse incentive against efficiency — being faster means earning less. Plaintiff firms work on contingency, taking a cut only when they win, so AI that lets them handle more cases at lower cost per case translates directly into revenue. Eve charges based on the number of clients a firm represents, which keeps incentives aligned.

The competitive dynamic AI is creating

Defense firms have traditionally used discovery requests to slow plaintiff attorneys down — a process that can take a plaintiff firm 48 hours to complete per request. Eve customers are returning those responses almost immediately, which Jay says is raising settlement values by eating into the billable hours the defense side was banking on. The plaintiff firm, which once showed up to court as one attorney against ten insurance company lawyers, can now respond at a pace that pressures the other side to settle earlier.

About 70% of personal injury cases are resolved before reaching a formal lawsuit, typically at the demand letter or mediation stage. Jay argues that faster, more capable plaintiff representation will push more cases toward earlier settlement rather than exponentially increasing courtroom litigation — though more marginal cases that previously weren't worth taking on will now get representation.

Access to justice and the lower end of the market

In labor and employment, plaintiff attorneys have historically required cases with at least $5,000 in potential fees before agreeing to represent someone. Eve is enabling firms to take on smaller or previously uneconomic cases, which Jay describes as effectively expanding access to legal representation down the income ladder.

Model architecture and data moat

Eve was among the first production applications on Claude 2, and Anthropic extended its context window specifically to accommodate the platform. Jay describes the data advantage not as proprietary case data but as capturing the intermediate workflows that don't exist anywhere publicly — the multi-step document chains a firm uses to produce a final demand letter, for instance, which can't be Googled. Eve lets firms teach the platform their internal workflows the way they would onboard a new associate, and it then executes those workflows consistently without the errors that erode lawyer trust in generic tools.

Jay argues that general-purpose model improvements — better reasoning, longer context windows, stronger agent functionality — directly expand how deep Eve can go into a legal workflow, so foundation model progress is additive rather than threatening to the product.

The longer-term structural bet

Jay's prediction is that leading plaintiff law firms will increasingly function like software companies — consolidating expertise in a narrow vertical, such as traumatic brain injuries, and serving every case in that category rather than being constrained by attorney headcount. The analogy he draws is to early AWS: tasks that previously required human labor become API calls, and marginal cost per new client drops toward zero. A single specialist attorney running a $100 million revenue firm is the extreme version of where he thinks this goes.