AI productivity is new demand before it is lower delivery cost

On October 28, the Thomson Reuters Institute published its Law Firm Rates Report 2027, which says AI is boosting law firm profitability mainly by generating new legal demand rather than by improving internal efficiency. The report also says record rate increases and strong client demand are sustaining growth, while firms need to measure AI’s impact, rethink pricing and staffing, and prepare for efficiency-driven disruption. The development arrives as Legaltech News reports that in-house teams are sending less work to outside counsel, leaving law firms to pass on AI-driven savings and examine new billing models. The immediate news is therefore not that AI has already displaced hourly firms, but that the market is beginning to separate demand growth from productivity growth.

That distinction matters because the billable hour makes a firm’s productivity gain economically awkward. If AI reduces the labor needed for a matter, an hourly firm either collects less for the same work, charges for time that was not worked, or must find more work to replace the lost hours. Fixed fees do not create that same arithmetic problem: lower delivery cost can become margin, a lower price, or both. The Thomson Reuters Institute report points to new demand as the current source of profitability, not to internal efficiency, while Legaltech News describes clients retaining more work and firms considering different billing models. Together, those accounts support the narrower position that AI productivity is not yet reliably translating into hourly-firm advantage.

Clio supplies the strongest objection to that reading. In a company release, Clio says its 2026 Legal Trends Report expects billable time to reach 42% of a lawyer’s workday, up from 39%, calling that the largest year-over-year increase. A rise in recorded billable time is plainly inconsistent with a simple story in which AI immediately shrinks every firm’s revenue base. But it does not show that AI shortened the same matters while firms preserved revenue without billing unworked time. It may instead reflect the new demand described by the Thomson Reuters Institute, or the continued strength of client demand and rate increases. The number is a challenge to an immediate collapse of hourly economics, not proof that hourly pricing captures AI productivity efficiently.

The reports establish a market tension, not a completed transfer of legal work. They show a profitability story still tied to demand, a workload story in which in-house teams do more with AI, and a pricing response that is beginning to move beyond hours. They do not establish that fixed-fee firms are already taking comparable recurring matters from hourly incumbents, that a particular firm has lost revenue on an AI-shortened matter, or that a firm’s fully loaded cost per completed matter has fallen. Those are the facts required for the stronger conclusion that fixed-fee delivery is already winning on price or margin. Until those facts appear, the claim that the billable hour is a liability remains directionally supported but commercially unproven.

The next confirming fact would be a firm or client report that ties a recurring category of work to all three parts of the shift: AI reduced the labor required, an hourly incumbent lost the work or revenue, and a fixed-fee provider delivered it at a lower price or cost. A future pricing report showing the fixed-fee share rising in contested work would strengthen the case. A report that hourly firms cut labor on AI-assisted matters while keeping revenue without billing unworked time would cut against it. We will treat the distinction between new demand and lower delivery cost as the central test: AI use alone says little about whether hourly billing is still competitive.