You have decided to train your teams on AI. Good move. But do it badly and what you get back is crossed arms, quiet eye-rolls, and adoption close to zero. Not because your teams are closed off. Because you touched a nerve. AI makes people anxious, and that anxiety has only grown over the past year. Fear does not get fixed with a slide deck.
Here is the picture that says it all. Roughly three in four knowledge workers already use AI at work, and only 39% of them have had any training from their employer. Your teams are not waiting for your permission. They are moving ahead with no frame. And a large share of them do not trust their company's AI strategy. The problem is not adoption. It is trust.
I train teams all year, from the executive committee to the front line. And what I see is that resistance almost never comes from the technology. It comes from the way it gets introduced. Change the approach and you change the result.
Fear number one: "AI is going to replace me"
It is the first thing your people think, even the ones who never say it out loud. "If AI does my job, what am I for?" The question is fair. And if you do not meet it head-on, you just let it grow under the surface.
Here is what I say at the start of every session. AI removes tasks, not jobs. The difference is everything. Nobody was hired to write meeting notes. You were hired for your judgement, your client relationships, your ability to solve problems. The meeting notes are plumbing. AI does the plumbing. You do the rest.
When people get that, they relax. And once they feel, concretely, how much time they get back, they never want to go back.
Do not train. Let them try.
The worst format for AI training is the lecture. Two hours of slides on what AI can do, generic examples, a quiz at the end. Everyone nods, nobody changes a thing the next day.
What works is the workshop. You take a real problem the team has. A real email to write, a real report to summarise, a real brief to produce. And you work it live with AI, in front of them, with them. They see the result. They touch it. Then they try it themselves.
In a session I ran for a leadership team, I asked each person to show up with a concrete problem. Within an hour, every one of them had solved theirs with AI. Three of them told me afterwards: "I did not think it was this simple." Simple brings people in. Complicated pushes them away.
Start with the willing, not the skeptics
You will not convince the skeptic first. And it is a strategic mistake to try. The skeptic needs proof, not arguments. And the best proof is what their own colleagues pull off.
Find three to five people on the team who are curious, who have already tried a tool, or who keep complaining about repetitive tasks. Train them first. Give them time to experiment. And ask them to share what they find in the team meeting. Not a formal presentation, just a "look what I did this week".
The skeptic will watch. And when they see a colleague turn out a report in twenty minutes instead of two hours, they will ask to learn. It always goes this way.
Three mistakes that kill adoption
First mistake: forcing one single tool on everyone. Every person has their own habits, needs, pace. Impose one tool on the whole team and you manufacture frustration. Let people explore. What counts is the result, not the tool.
Second mistake: measuring adoption by the number of logins. Someone can open an AI tool every day and get nothing useful out of it. What counts is the time saved, the quality of the work produced, the satisfaction of the team. Measure that.
Third mistake: not training the managers. Managers are under real strain right now, caught in the middle, squeezed between the AI mandates coming down from the top and the resistance coming up from their teams. If the direct manager cannot use AI, they will not know how to champion it. Worse, they may see it as a threat to their authority. Train the managers first, or at the same time as the teams. Never after.
The manager's role in adoption
A manager who uses AI in front of their team legitimises the practice. A manager who says "I prefer to do it the old way" kills it. It really is that simple.
I tell managers to share their own uses in the team meeting. "This week I used AI to prep our budget review and saved myself forty minutes." It takes the mystery out of the tool, it shows it is not reserved for the tech crowd, and it creates an implicit permission for everyone else.
An operations director I work with set up a simple ritual. Every Monday, at the start of the team meeting, someone shares one "AI trick" they discovered. Two minutes, no more. Within three months, the whole team was on board. No mandate, no extra training. Just positive contagion.
The all-or-nothing trap
Plenty of leaders want to roll AI out across the whole company in a single quarter. One plan, one budget, one tool, one timeline. On paper it looks clean. In practice it is the surest way to trigger a revolt.
Adoption happens in circles. You start with a pilot team. You validate the uses. You document the gains. Then you widen to a second circle, then a third. Each circle benefits from the experience of the last. And the skeptics in circle three are already sold by the results of circles one and two.
This approach takes a little longer up front. But it produces durable adoption, not a flash in the pan that burns out in three months.
What you have to protect
Training on AI also means training on its limits. Say it plainly from the start. AI can be wrong. It sometimes makes up numbers. It does not replace verification. And above all, confidential data does not go into just any tool.
Set simple rules. Which data can be used, which cannot. Which tools the company has approved. Who to contact when in doubt. And remember that the EU AI Act now requires a baseline level of AI literacy for anyone using an AI system at work. These rules do not slow adoption down. They make it safe. And they show your teams you take the subject seriously, not as a gadget.
Frequently asked questions
How do you train your teams on AI without creating resistance?
By starting from their problems, not the technology. Show AI on a task they hate doing. The time it saves is what wins them over. And above all, do not make AI mandatory at the start. Let the volunteers lead by example.
Should you train everyone at the same time?
No. Start with a small group of volunteers, three to five people. They test, they share, they create the pull. Skeptics move when they see the results of others. Forcing everyone at once creates resistance.
Will AI cut jobs on my team?
AI removes tasks, not jobs. What it eliminates is the time lost on repetitive work. What it frees up is time for thinking, client relationships, creativity. Companies that train their teams on AI do not lay people off. They level up their skills.
How long does it take to train a team on AI?
For the basics, a single day is enough. To lock in the reflexes and weave it into daily work, count on about a month of follow-up. What matters is not the length of the training, but the follow-up that turns a one-off try into a habit.
Want to train your teams on AI without the pushback? AI x Leaders works with companies through hands-on workshop formats, tailored to your sector and your team.
Let's talk about your project →
