Why Your Team Isn’t Using AI
and How to Change That Without the Hassle
The three forms of resistance every leader encounters, and the levers that actually work
Thursday afternoon, three weeks after the AI kickoff.
You walk through the office and see two employees still putting together proposals by hand. A third has the AI tool open but is typing their own text and copying it in “just to see if it’s right.” The enthusiastic colleague who sat at the front during the training is the only one genuinely using it.
You’ve invested. You organised a training. You bought licences.
And yet.
This is not an exception. It is the standard. McKinsey research shows that in most AI implementations in small and medium-sized businesses, active adoption drops back to less than 25% of the team after six weeks. Not because the tools are poor. Not because employees are lazy. But because technological change works differently from what most leaders expect.
It’s Almost Never a Technical Problem
When adoption stalls, the first reaction of many directors is: organise better training, create a manual, make the tool more accessible.
That sometimes helps. But it rarely addresses the root cause.
The root cause is human. People don’t change their way of working because a director decides so or because a tool is available. They change when they understand for themselves why it’s better — and when they feel safe enough to learn without being judged for their mistakes.
That sounds obvious. In practice, this is almost always where things go wrong.
The Three Forms of Resistance
Resistance to AI in the workplace is not one thing. It is at least three different things, and each requires a different approach.
Resistance 1: Fear of Replacement
“If I use this well, I’ll make myself redundant.”
This is the most fundamental barrier. Employees who have delivered value for years through specific skills — writing proposals, producing reports, analysing data — experience AI as a threat to that value.
They will rarely say this out loud. But you recognise it in subtle behaviour: not keeping up with training sessions, finding reasons why “it’s different here for us,” or pointedly doing things by hand when they could easily do otherwise.
Resistance 2: Perfectionism Paralysis
“I don’t know if I’m using it correctly.”
This is the most common barrier among motivated, careful employees. They want to do things properly. They see that AI output isn’t always perfect. They’re afraid of sending or delivering something that contains “AI errors” — and damaging their reputation in the process.
The result: they open the tool, aren’t satisfied with the first output, close it again, and reach for the familiar document.
Resistance 3: The Power of Habit
“I’ve been doing it this way for ten years. Why would I change?”
This isn’t cynicism. This is cognitive economy. People who know their work well have developed automated routines. Those routines require little mental energy. Learning a new tool — even one that is ultimately faster — initially costs more energy than the familiar approach.
As long as the benefits are not visible and personal, the rational choice is: don’t change.
What Doesn’t Work
Three classic implementation mistakes that increase resistance rather than reduce it.
Imposing from the top down without context
“From next month, we’ll use AI for all client communications.”
Without explaining the why, without space for questions, without acknowledging what employees are feeling. That creates uncertainty and resistance in exactly the people you need most.
One big training session, then nothing
A two-hour kickoff with an external trainer. Enthusiasm. Notes. And then: nothing. No follow-up, no shared experiences, no opportunity to ask questions when you run into something a week later. Learning doesn’t work in a single session. Especially not when you need to change something you’ve done the same way for ten years.
Positioning AI as a cost-cutting measure
This is the most damaging mistake. If employees — even indirectly, even through the grapevine — receive the signal that AI means fewer people are needed, the resistance will never go away. It is the fastest way to ensure that the people you need most to make AI work are precisely the people who will undermine it. Not deliberately. But effectively.
Three Levers That Do Work
Lever 1: Early Adopters as Ambassadors
In every organisation, there are people who embrace new things. Who were already experimenting before there was an official policy. Who were enthusiastic during the kickoff and actually kept going afterwards.
Most directors leave these people out of the adoption campaign. They communicate themselves — by email, in meetings, via announcements. But employees don’t listen to the director when it comes to “is this worth my time?” They listen to colleagues.
Give early adopters the space and the platform to share their experiences. Not as a formal presentation, but informally. “Just show them what you do with it.” A colleague who says “this saved me 45 minutes yesterday on that client letter” does more than ten emails from the director.
Lever 2: Small, Visible Wins in the First Week
Not the big transformation. Not the explanation of the strategic vision. Not the story about where the company wants to be in three years.
One task. One concrete time saving. This week.
The most effective AI adoption doesn’t begin with a plan but with a moment when someone thinks: “Oh. This actually works.” That moment needs to come as quickly as possible, and it needs to be personal. Not “the company saves time” but “I saved half an hour today.” Help employees find that moment. Ask them: what is the most time-consuming, repetitive task in your work? Start there. Not with the most complex, not with the most visible — with the most frustrating.
Lever 3: Psychological Safety
This is the lever with the greatest impact and the lowest cost.
Say — explicitly, as a director — that it’s fine to fail. That you don’t expect everyone to get it right in week one. That output that isn’t quite right isn’t a problem, as long as a human has reviewed it before it goes out.
That sounds simple. But the difference between a culture where people dare to experiment and a culture where people wait until they’re sure it’s right is, in most cases, one or two statements from the director. The tone you set determines the adoption speed of your entire team. If you show that you’re learning too — that you’ve had poor AI output yourself and can laugh about it — you give the rest of the team permission to do the same.
The Director’s Role: What You Should and Shouldn’t Do
What you do:
- Set boundaries (the three rules)
- Celebrate successes, including the small ones
- Give ownership to people who can handle it
- Be visible in your own learning process
What you don’t do:
- Control and validate every decision
- Link AI to cost-cutting objectives
- Expect adoption to happen by itself after one training session
- Wait until everyone is ready — that moment never comes
The director who successfully lands AI in their organisation is not the director who knows the most about AI. It is the director who creates the right conditions for their people to discover it.
What Now?
This week, identify one employee who is already working enthusiastically with AI — even informally, even without an official brief.
Ask them: “Would you be willing to show three colleagues what you do with it?”
No presentation. No meeting. Just showing what you do.
That is the cheapest, most effective adoption campaign you can run.
Want to know how to build buy-in in your specific team? At LinkLeaders, we combine the technical approach with the human side — because no AI plan works without people.
Call +31 (0)182 570927 or email contact@linkleaders.nl.