Getting your team comfortable with AI tools (without the eye-rolling)

Introducing AI to a team is more about people than technology. Here is how to do it without the eye-rolling, the fear or the forced enthusiasm.

The technical side of adding AI to a small business is usually the easy part. The awkward part is people. Some team members are quietly worried about their jobs, some think it is a gimmick, and some have already tried a tool once, got a rubbish answer and written the whole thing off. Getting a team comfortable is mostly about honesty and small steps.

Start with a real annoyance, not a grand plan

Nobody gets excited about “our AI strategy”. People do get interested when a tool quietly removes a job they hate. So start there. Find one genuinely tedious task and show how a tool shortens it. A single visible win does more for buy-in than any all-staff presentation.

Crucially, let the person who does that task be involved in choosing it. Tools imposed from above get resented. Tools that solve a problem someone actually mentioned get adopted.

Be honest about what it can’t do

Overselling is the fastest way to lose a team. If you tell people a tool is brilliant and their first attempt produces nonsense, they stop trusting both the tool and you. It works far better to say plainly: this is good at first drafts and summaries, it makes things up sometimes, always check it. That honesty gives people permission to be sceptical, which paradoxically makes them more willing to try it.

Address the job-security worry directly

The eye-rolling is often nerves in disguise. If people think a tool is there to replace them, they will resist it, and reasonably so. Be clear about the actual intent, which in most small businesses is to take the dull admin off people’s plates so they can do the work that needs a human. If that is genuinely the plan, say it out loud. If it isn’t, they will find out anyway, and no amount of training will fix that.

Make it easy and safe to experiment

People get comfortable by playing, not by being trained at. A few things help:

• Give clear rules on what not to put in. No customer personal data, no passwords, no anything you wouldn’t email to a stranger, unless you are using a properly set-up business tool. This overlaps with basic data safety and is worth spelling out early.
• Share prompts that worked. A small shared list of “try this” examples beats a formal course.
• Let people opt in. Enthusiasts first, sceptics later once they see it working. Forcing everyone at once guarantees resistance.
• Keep expectations low at the start. “See if it helps” is a better brief than “use this for everything”.

Give it a little structure

Total free-for-all and rigid mandates both fail. A light framework works best: which tools are approved, what they are and aren’t for, and where to ask if unsure. If you want a sense of how a measured, un-hyped rollout tends to go, our approach to how we work is built around exactly this sort of small, honest step. And when a tool proves its worth, a purpose-built assistant shaped around your actual workflow usually beats a generic one, because it fits how the team already does things.

Go at the pace of the most sceptical sensible person on the team, and let results do the persuading.