AI and the billable hour: a reckoning for legal
When AI does ten hours of legal work in one, the billable hour stops measuring value and starts penalising efficiency. The pressure falls on how law firms price their work.
TL;DR
- In 2025, top United States law firm rates passed 1,000 dollars an hour, with some reaching 2,000, even as AI compresses ten-hour tasks into one.
- About 90 percent of legal revenue still flows through hourly billing, a model essentially unchanged since the 1950s.
- The pricing model is what is under threat, as value moves from hours logged to judgement applied.
Thomson Reuters frames the tension precisely in The 2,000 hour problem. In 2025, standard rates at the largest United States firms crossed 1,000 dollars an hour for the first time, with some partners reaching 2,000. At the same time, work that once took ten hours of research, review, and drafting can now, in principle, be done in one.
The billable hour rewards the time a task takes, and AI's entire purpose is to make that time disappear.
A model in collision
The numbers describe a model in collision with the tools now entering it. Law firms increased technology spending by about 9.7 percent in 2025, the fastest real growth on record, racing to deploy AI that synthesises case law, drafts contracts, and reviews documents at speed. Yet around 90 percent of legal dollars still move through hourly arrangements that have barely changed since the 1950s.
| The pressure point | Detail |
|---|---|
| Rates | Top firm rates passed 1,000 dollars an hour in 2025, some at 2,000 |
| Efficiency | Routine ten-hour tasks now achievable in roughly one |
| Investment | Legal tech spending up about 9.7 percent in 2025 |
| Inertia | About 90 percent of revenue still billed by the hour |
The conflict is structural. If a firm bills by the hour and AI cuts the hours, revenue falls even as client value holds or rises. Several commentators, and a growing number of ethics opinions, argue it is hard to justify charging full hourly price for work that no longer takes the time.
Repricing around judgement
The billable hour is under pressure because it measures the wrong thing. It prices input, and AI attacks input. The firms that thrive will be those that reprice around output and judgement: fixed fees, subscriptions, and value-based arrangements for the routine work AI now accelerates, and premium pricing for the high-stakes strategy and advocacy AI cannot do.
This is a professional-services story as much as a legal one. Any business that bills time and sells expertise, including consultancies and agencies, faces the same question. When the cost of producing the deliverable collapses, you can defend the old price or reprice around the value. Only one of those is stable.
Where to begin
- Separate routine, automatable work from high-judgement work, and price them on different logics.
- Move repetitive work toward fixed or subscription fees, so efficiency rewards the firm instead of cutting its revenue.
- Reframe the client conversation around outcomes and risk, ahead of clients forcing the change on their own terms.
Related reading
- Will AI mean the death of the billable hour? (Clio)
- Working the jagged frontier: judgement as the core AI skill
Source: The $2,000 hour problem: when AI efficiency collides with billable time, Thomson Reuters.