Guides / Team adoption

Guide · Adoption

Getting Your Team Actually Using AI

01Why "we bought the licences" isn't a plan

Here's the pattern we see: a business signs up for an AI tool, announces it at a team meeting, and three months later two people use it occasionally and everyone else has quietly gone back to how they always worked. The subscription renews. Nothing changed.

The tool was never the problem. Nobody resists AI because the software is hard to open — they resist because nobody answered three unspoken questions: Am I allowed? Am I safe? What's in it for me? Adoption is the work of answering those three questions, and it costs more attention than money.

02Deal with the elephant first

The biggest unspoken question is the obvious one: "is this thing being brought in to replace me?" If you don't answer it, your team will answer it for themselves — and then politely, invisibly, make sure the AI never looks too useful.

So say it out loud, and mean it. In a small business the honest answer is almost always that AI changes tasks, not jobs: it takes the typing-up, the chasing, the formatting — the parts of the job nobody fought to keep — and leaves the customer relationships, the judgement and the craft. The pitch to your team isn't "be more productive"; it's "let's bin the boring bits of your week." That's a pitch people can get behind.

One caution: don't promise what you can't keep. If efficiency gains are part of a bigger plan, be straight about it. Teams forgive honest uncertainty; they don't forgive being managed.

03Give permission — with rules

Here's something most owners don't realise: some of your team are probably using AI already, on their personal phones, without telling you. Researchers call it "shadow AI", and it's the worst of both worlds — you carry the data risk without getting the open benefit.

The fix is to legitimise it. A one-page policy that says yes, use it — here's how beats both silence and a ban. The four house rules from our data safety guide are designed to be exactly that page: no personal data in free tools, anonymise where you can, sensitive material only in business-tier tools, and a human checks anything that leaves the building.

Permission with rules feels like trust. Silence feels like a trap. Bans just push it back onto personal phones.

04Start with pain, not features

Don't train people on what the tool can do; ask each person what they'd most like to never do again. The answer — typing up site notes, the Monday report, replying to the same five questions — is their starting point. It's the repetitive-task test applied per person.

This matters because the first win has to be felt personally. "AI saved the company time" moves nobody. "I got Friday afternoon back" gets discussed at lunch — and word of mouth inside a small team is the most powerful adoption tool there is.

05Make space, and find your champion

Two structural moves that cost little and change everything:

• Give it sanctioned time. An hour a week, on the clock, to try things — plus five minutes in the team meeting for "anything AI saved you time on this week?" Shared prompts beat any manual. If learning only happens in stolen moments, it loses to the day job every time.
• Anoint a champion. Every team has one person who's already curious — often not the most senior or most technical. Give them the role officially: they collect what works, help the hesitant, and feed problems back to you. Enthusiasm spreads peer-to-peer far faster than top-down.

06Lead from the front

If the team never sees you use it, they'll conclude it's homework rather than how-we-work-now. Use it visibly. Share your own wins — and your failures, which are honestly more useful: "I asked it for the figures and it confidently made them up, here's what I do differently now" teaches the checking habit better than any policy document.

07Measure lightly, keep what works

Skip the dashboards. One question a month per person is enough: roughly how many hours did it save you, and on what? Keep what's saving time, drop what isn't, and let the wins compound. Most teams that follow the steps above find a rhythm inside a couple of months — not because they were made to, but because the boring bits of the week genuinely shrank.