If you've been wondering whether your legal team is behind on AI, I'd look at the question a little differently.
Legal AI isn't a race. It's closer to a marathon, so pace yourself and think long term. It's also several things at once: tools, training, data, measurement and oversight. A team can be well along on some and just starting on others. At Execo, our AI Center of Excellence built an in-house tool that gives us visibility and control across finance, sales and delivery.
So it doesn’t really help to compare who is ahead and who is lagging behind. Instead, I’d recommend you identify and recognize where your team is strong, where it’s still building, and what is coming next.
FTI Consulting and Relativity's 2026 General Counsel Report, a survey of more than 200 general counsel at large companies, found that 87% report AI use in their teams, up from 44%. Only 39% see AI as a strategic priority, and 53% have a formal technology roadmap, up from 25%. The 87% counts reported use, not a legal tool that's rolled out and working. Three measures at three different levels suggest progress is uneven even among large legal departments.
"Further along" isn't one measurement. A team can be well along on tools and still building its data, or the other way round. Where you're working things out may be where another team is comfortable, and the reverse.
The progress worth tracking is an outcome you already care about: turnaround time on a standard contract, outside-counsel hours on one matter type, or work your lawyers have stopped doing by hand. If one has moved, that's progress, whatever anyone else is doing. Using AI case by case, with lawyers keeping the judgment calls, is also a sound way to work for some kinds of work. It's a decision, not a gap.
Wherever you are, here's what each stage calls for.
Still choosing. Pick one kind of work and say what better looks like. Then find out where you are today: take last month's NDAs and note, for each, the day it reached legal and the day legal sent it back, then average the gap. If you keep a tracker, the dates are already there. If not, a handful of recent NDAs from your inbox or shared drive will give you a number to compare against later.
Trying it on real work. Help people get comfortable with work they already do. In my view, a tool that's easy to pick up is more likely to stay in use. One legal leader I spoke with described daily users growing confident while occasional users relearn the tool each time. Pair an occasional user with a daily user on one real matter, and let the occasional user drive. If no one uses it daily yet, pair two people who've each tried it once.
In everyday use. Choose what to measure and how quickly to widen use. Switch features on in stages, or stagger access if the tool can't. Pick one metric you can pull from records you already have, such as outside-counsel hours from invoices.
What your answers suggest:
None of this is a grade, and a team can answer differently for different kinds of work.
Progress in legal AI comes in different places at different times.
The comparison that helps is your own work against a year ago. Is any of it faster, cheaper or lower-risk? If yes, you're further along than a count of licenses would say. If not yet, you know which stage you're in and what to do next. And if you're staying with case-by-case use on purpose, that's a fair decision.
We can talk about it more over coffee so we can compare notes. Feel free to reach out!