How do you lead your team through AI change from the middle?

The position you are in

The decision to buy AI was made above you. AI arrived with a launch email and a target. Your team looks to you for what it means for their jobs, and the people above you look to you for results. You are expected to lead a change you did not design, with AI you may not have had time to learn yourself.

That position is common, and it is hard. In BCG's 2025 AI at Work survey, only about a quarter of frontline employees said their leaders gave them enough guidance on AI (BCG, June 2025). Your team will look to you first, so that guidance has to come from you.

Start with one task, on real work

A demo of everything AI can do is hard to act on. Start with a single task your team already does every week, so each person knows what to try on Monday morning.

Picture a team of twelve in a marketing operations group. In week one, each person uses AI to draft the first version of the weekly status note they write anyway. It takes twenty minutes. Two people find it saves them half an hour. One person thinks the draft is worse than their own and says so. All three reactions are useful. Next week, the two who found the time show the others what they did differently. The person who pushed back explains what the draft got wrong, and the team learns where AI needs checking.

That is the whole method, repeated. Our article on which workflows to point AI at first helps you choose the task.

Treat pushback as information

When an experienced person holds back from AI, they are protecting something: the quality of the work, their expertise, or their place on the team. Ask what they are seeing before deciding they need convincing. They may have spotted a real limit the enthusiasts missed. If the worry is about their role, say plainly what the change means for them, as far as you know it. Our page on why employees resist AI goes deeper.

There is a second cost. In BCG's 2026 survey, 47% of workers said they spend more time managing and directing AI than doing the work itself (BCG, June 2026). If your team says AI is creating work, believe them, and look at the step where it happens.

Decide where the saved time goes

When AI saves your team an hour, that hour fills with something. If nobody decides what, it fills with more of the same work, and the people above you see no change. Agree with your team, in writing, what the first saved hours are for: a backlog item, a piece of work that never gets time, or a skill someone wants to build. Then tell your own manager. Our article AI saved your team hours. Where did you send them? covers this in detail.

In the language of the AI Profitability Gap™, the first steps are Skill and the last one is Reinvestment. Both sit with you.

Show your results upward

Count how many people on your team use AI on a recurring task today. Count again in eight weeks. Write down where the saved time went. That gives the people above you a before-and-after they can use, and it gives you a stronger case for the next round of support.

How we help a leader in the middle

The AI Profit Sprint is built for this position. It is eight weeks for one leader and a team of up to 20, with about 20 minutes a week for each person. Each week your team gets a task card set on work they already do. You get two coaching sessions with Rob Cannon, PhD, an organizational psychologist, and at the end a team playbook and a before-and-after readout written so you can take it to the person above you. It costs $8,500. If you want to start smaller, the AI Profit Workshop is a founder-led half-day for you and up to six people, and its $2,999 fee is credited in full toward the Sprint if you start within six months.