[Wafer: response was truncated before the model finished its internal reasoning. Increase max_tokens, or disable thinking on this model (e.g. chat_template_kwargs.enable_thinking=false), then retry.]

Law firms boosted technology spending 9.7% and knowledge-management investment 37.2% since 2021, yet BARBRI Professional Education research published August 6 finds most deployed AI faster than they built ways to measure whether lawyers actually work differently. The gap matters because roughly 90% of legal fees still flow through hourly billing, according to the Thomson Reuters and Georgetown Law State of the US Legal Market report.

The research draws on interviews with ten leaders across nine firms. Nearly every firm has made real progress getting tools into lawyers' hands. Almost none has built the organisational plumbing to sustain training, drive adoption, or produce a number that would survive contact with a sceptical managing partner.

One former learning-and-development director summed up the internal reality as "feast or famine": either the innovation, knowledge-management and L&D teams are all elbowing for the same territory, or nobody owns it at all.

The measurement gap

Firms bought Harvey seats and CoCounsel licences at speed, then discovered that measuring behaviour change requires someone whose actual job is to measure it. When three teams share the tools, they can easily trip over one another. Either way, the partnership is paying for a capability it cannot yet describe to a client.

The hourly billing tension

The Thomson Reuters and Georgetown Law report found that around 90% of legal fees still flow through standard hourly arrangements, the same structure that has governed the profession since the 1950s. Set that against a spending surge on tools built to compress hours into minutes, and you get what the report calls an "absurd tension."

The maths only works if firms can push through rate rises steep enough to outrun their own efficiency. Clients are not thrilled at the idea of handing every productivity gain straight to partner profit, and general counsel have started signalling spending pullbacks. Rising associate costs make the squeeze sharper still.

The unbilled work opportunity

For contrarians, the bubble-bursting narrative is only half right. Associate realisation rates already sit around 85.6%, meaning a meaningful slice of junior work never gets collected. AI can absorb the unbilled, unglamorous portion of the job without touching a dollar of revenue that was ever going to arrive.

That explains why elite firms are layering AI on top of headcount rather than cutting, with average midsized and second-hundred firms growing lawyer numbers by more than 8% since January 2023. The smart play is not to spend the most but to automate work that clients were not paying for anyway.

The downturn that didn't arrive

The predicted 2026 downturn did not arrive in the manner many expected. Demand kept climbing through the first half of 2026, with Wells Fargo clocking a 13.1% first-quarter revenue jump across the Am Law 200 and standard rates still rising into double digits.

Thomson Reuters' own first-quarter index landed at 55, sitting dead on its twenty-year average, despite some of the strongest rate and demand inputs the market has ever recorded. Strong inputs, ordinary output.

Collections have begun to drag, with the cycle slowing more than 3% as unbilled work stacks up at the top of the market. The report called this the first drops of rain.

The average firm booked 13% profit growth in 2025. That is a genuinely good year built on genuinely unstable ground. The firm that gets caught out will be the one that spent the money, briefed the press, and still cannot show the client a number.

Why this matters for legal professionals

For lawyers, AI adoption is not the differentiator - the measurement and billing infrastructure around it is. Firms that can show clients a number for what AI actually saves will hold pricing power. Those that cannot will face the double squeeze of rate pressure and collections drag. Individual lawyers should track how their firm measures AI's impact, because that metric will determine how their own productivity is valued at the next partnership review.