Why your team isn't using AI every day, and how to fix it

Most teams have tried AI. Far fewer are using AI every day, and that distinction is the whole ballgame. Gallup's late-2025 data found only about 12% of US employees use AI daily, while roughly half don't use it for work at all. The common picture isn't a team that rejected AI; it's a team that picked it up, used it once or twice, and drifted back to old habits. People get stuck in the explorer phase, where AI is an occasional novelty rather than a daily tool, and that's not a deployment problem you fix with another license. It's a habit problem, and habits are built differently than rollouts.

This matters because the value isn't spread evenly across occasional and habitual users. It's lopsided. So before you can fix the frequency gap, it helps to see why frequency is the thing worth fixing at all.

OpenAI's own analysis of workplace usage found a roughly sixfold gap between power users and the median employee at the same company, and framed the divide exactly right: not between those who have access and those who don't, but between those who've made AI a daily habit and those for whom it stays an occasional novelty. The people at the top of that distribution, the ones using it every day, report saving more than ten hours a week. The occasional users save almost nothing, because a tool you touch once a fortnight never becomes faster than your old way of doing the work. That's why daily use is worth caring about: the return on the whole AI spend is concentrated in the habitual users, so moving people from now-and-then to every day is where the payoff lives.

Here's the fair objection, and it's a good one: isn't chasing daily use just surveillance with extra steps? If you mandate frequency and track it per person, aren't you measuring a vanity metric and breeding resentment? Partly, yes, and that's the trap to avoid. When daily use becomes a target you police, people game it, opening the tool to hit the number without changing how they work, and the heavy-handed version backfires: a meaningful share of employees required to use AI report it didn't help, and some quietly undermine the effort. The lesson isn't that frequency doesn't matter; it's that you can't compel it. Frequency is a result you engineer by making AI genuinely useful, not a behavior you force or a dashboard you wave at people. Read your usage numbers as a health signal for yourself, never as a scoreboard for individuals, and the whole exercise stays on the right side of the line.

Daily use comes from a cue, not willpower. People don't build a habit by remembering to go use the AI tool; they build it when AI is attached to a task they already do, so the work itself is the trigger. Getting unstuck, summarizing a long thread, drafting the first version, these become the cues that pull someone back to AI without a reminder. It's why Microsoft saw daily active use of its assistant jump roughly tenfold after it was built directly into the apps people already work in, instead of sitting on a separate site. The practical move for a team is to attach AI to a recurring, real task in each role, get an easy first win there, and let the work keep triggering the habit. That's how a trial turns into a routine, and it's the engine of real AI adoption.

The single biggest lever on whether a team uses AI every day is the manager. Gallup found that employees whose manager actively supports AI use are more than twice as likely to use it regularly, and far more likely to feel it helps them do their best work, yet only about 28% of employees say their manager actively supports the team's AI use. That gap is the opportunity. The manager who models AI in the open, asks the team what they tried this week, celebrates a real win, and helps whoever is stuck does more for daily use than any mandate, because they're reinforcing a habit instead of policing a metric. Pair that with a champion on the team who shows peers AI on their own tasks, and frequency climbs on its own. Daily use, in the end, is something you earn by making AI worth reaching for, not something you count on a screen. Keep it an outcome of good team habits, and it sticks.