How to tell if your team is actually using the AI tools you pay for
You bought the seats. ChatGPT, Copilot, maybe both. A few people were excited at the kickoff. Three months later you're paying every month and you have no real idea whether anyone uses it.
You're not the only one. In Gallup's August 2025 survey of 23,068 US workers, 45% said they use AI at work at least a few times a year. Only 10% use it every day (Gallup). Most teams sit somewhere between "tried it once" and "use it for everything."
Here's how to find out where yours is.
Don't judge it by the usage report
Admin dashboards count logins and messages. They can't tell you whether the tool saved anyone an hour. Someone who opens Copilot once a week to tidy an email counts the same as someone who drafted a whole client proposal with it.
The report also misses the tools you didn't buy. In Microsoft and LinkedIn's 2024 survey of 31,000 workers, 78% of people using AI at work brought their own tools (Microsoft). Your dashboard can look quiet while your team uses a free app you've never heard of.
Use the report for one number: how many seats nobody has touched. That's money you can save or redirect today.
Ask, but not in a meeting
The quickest way to find out is to ask. The worst way is a team meeting where you say, "So, is everyone using the AI?" Nobody wants to tell the person who paid for it that they haven't opened it.
Three things make the question work:
- Keep it anonymous. WeeklyTake never shows names, never shows anyone's individual answers, and hides results until at least three people have answered.
- Keep it to one question. "How is the AI assistant working for you this week?" with three answers: Saving me real time, Still figuring it out, Haven't really used it.
- Ask at the right moment. Right after a training, or a week after someone first gets access. Not on a random Friday three months in.
The reason matters more than the number
A low score tells you something's wrong. The reason tells you what to do. When Gallup asked workers what holds them back, the most common answer was that they couldn't see how AI fit their job (16%). Lack of training came lower, at 11% (Gallup).
Each reason has a different fix:
- "I don't know what to use it for." Show two real examples from your own business at the next team meeting.
- "I tried it and the answer was wrong." Share a prompt that works well for that task, and say that checking the output is part of the job.
- "I'm not sure I'm allowed to." Write one short rule about what's fine to paste in and what isn't.
- "I don't have time." Give people 30 minutes this week to try it on a real piece of work.
How to measure AI adoption in a small business
One number won't tell you much. Four simple ones, tracked every week, will.
| What to track | Where it comes from | What it tells you |
|---|---|---|
| Seats nobody has opened | Your admin usage report | Money you can save or move |
| Share of the team using it weekly | One question: "How did AI fit into your work this week?" | Whether it's turning into a habit |
| The most common reason people are stuck | One question: "What got in the way?" | What to fix next |
| How many people answered | Your own count each week | A rough sign of whether people think answering is worth it. Busy weeks and time off move it too. |
Write the four numbers down once a week. One line in a spreadsheet is enough. In the first two months the direction matters more than any single week. Weekly use going up is a good sign. If the top blocker changes, check that the old one was actually fixed and people didn't just stop mentioning it. The same blocker three weeks running is worth a closer look.
Keep all four at team level. Measuring adoption person by person turns a check-in into surveillance, and people stop telling you the truth.
You are the biggest lever
Gallup found that employees whose manager actively supports their AI use are about twice as likely to use it a few times a week or more. Yet only 28% strongly agree that their manager does (Gallup).
Support isn't a speech at the kickoff. It's asking, doing something with the answer, and telling your team what you did.
A habit that takes about 15 minutes a week
- Monday: ask one question.
- Thursday: read the answers. Look for the most common reason people are stuck.
- Friday: tell your team one thing you're changing because of what they said.
Do that for a month and you'll know more about your AI rollout than any dashboard can tell you. That's the routine WeeklyTake runs for you, one email at a time.