Measuring AI adoption is the step most firms skip, and research reported across the trade press on 6 August puts the problem in a single line. Firms know who switched a tool on. Almost none know whether it changed the way anyone works. The study comes from BARBRI, an American legal education business, and rests on interviews with ten learning, knowledge and innovation leaders at firms ranging from the largest American partnerships to newer technology-led practices. That is a world away from a firm of eight in Sale, and the gap it describes is wider here, not narrower.
What the research found
Asked to mark their own technology rollouts, the firms gave themselves a middling seven out of ten. None of them had built a framework saying what competent AI use looks like at each stage of a lawyer's development, so nobody can tell a trainee what good looks like beyond keep an eye on it. The working relationship between the people who run training and the people who run innovation was described as fragile even where it had improved. Underneath all of it sits the billable hour, which the report calls a barrier that training on its own does not fix. A tool turning three hours of reading into twenty minutes takes money off the bill unless the firm prices the work another way.
Read that list and the pattern is administrative rather than technical. The software does what it says. What no firm had was a person whose job is to ask whether the work is being done differently, and a number that answers the question.
Why the gap is wider in a small firm
A large firm employs an innovation team and a learning team and still cannot answer the question. A firm of eight employs neither, so the question belongs to nobody and never gets asked. The licence renews on the anniversary, everyone says the tool is useful, and the partner who signed it has nothing to put against the invoice beyond a general sense that things feel quicker.
The second cost is one you cannot see without looking. Adoption in a small firm is rarely even. Two fee earners use the tool every day, one uses it for a single task and the rest opened it in the first week and went back to what they knew. That is not resistance, it is a training problem, a workload problem or a tool that does not suit the work they do. All three have a fix. None of them can be fixed while the firm counts licences and calls it adoption. There is a regulatory edge to this as well, because supervision means knowing who is using what, and a firm that cannot say who uses a tool cannot show it supervises the use.
What to count instead
Start with one task rather than the whole firm. Pick something you do often and that a tool is meant to help with, such as first drafts of a particular letter or the first read of a bundle. Write down how it works today and roughly how long it takes, because a baseline taken after the tool arrives is worth nothing.
Then count three things for ninety days. How many matters used the tool on that task, which tells you whether it reached the work at all. The time from instruction to a usable first draft, which is the claim the tool is sold on. How often the supervising fee earner sent the draft back, which tells you whether speed came at the cost of quality. Ask the people who stopped using it what stopped them, and write down the answers rather than trusting a corridor conversation. At ninety days decide one of three things, being widen it, fix the training, or drop it at renewal.
None of this needs software or a project. It needs an hour to set up, a spreadsheet, and somebody named as responsible for looking at it.
BARBRI publishes the research as Driving Change and ROI, Uniting Three Teams to Lead Law Firms into the Future, with the summary readable on the page and the full report released once you give your details.
If your firm has tools in place and no answer to whether the work changed, setting up that measurement is a short piece of work and it pays for itself at the next renewal: talk it through with us.