Your most experienced, highest-performing people are the ones pushing back hardest on AI adoption. And you’re probably treating this as a problem to solve.

It isn’t. Or rather, it is, but not the way you think.

The resistance you’re seeing from your best people isn’t ignorance, stubbornness or technophobia. It’s a perfectly rational response rooted in how human brains evaluate risk. These are the people whose existing methods have demonstrably worked. They’ve built careers, reputations and results using approaches they’ve refined over years or decades. When you ask them to change, you’re asking them to gamble everything that’s proven on something that hasn’t.

Understanding why this resistance is rational is the first step to actually overcoming it.

The psychology behind rational resistance

I’ve spent years researching the cognitive biases that affect creativity and innovation, and the pattern that shows up most reliably in AI adoption is this: the people with the most to lose resist the most. Not because they’re the least capable, but because the very success that makes them valuable has also made them psychologically invested in the methods that produced that success.

Let’s break down exactly what’s happening.

Loss aversion: losing feels twice as bad as winning feels good

The foundational research here comes from Amos Tversky and Daniel Kahneman’s work on prospect theory. They found that losing something you already have feels about twice as significant as gaining something of the same value.

Think about what this means for your senior sales director who’s closing deals at 30% above target. She has a system. It works for her, and it is what made her successful so far. You (someone who in her eyes has not achieved her level of success or experience in her field) are asking her to adopt AI tools that might make her more productive, or on the other hand might disrupt the workflow she’s perfected. The potential gain (maybe 15% more efficiency) doesn’t feel anywhere near as significant as the potential loss (her proven system, her sense of mastery, her identity as a success in her role, and possibly even her edge over colleagues).

As I’ve written about loss aversion’s impact on innovation, this bias is so strong that it’s been observed in studies with capuchin monkeys. It’s not a cultural artifact. It’s hardwired.

In many companies, when someone takes a bet on new technology and it pays off, they get a congratulatory handshake. When it doesn’t pay off, they get a very different conversation at their annual review, in their eyes due to something that they didn’t even want. Your best performers know this calculus intimately.

The status quo bias: when “good enough” wins by default

Closely related to loss aversion is the status quo bias, our emotional preference for the current state of affairs. Research has shown that our brain actually needs to work harder and use different pathways when making difficult decisions compared to easy ones. The brain literally prefers the default option because it requires less processing.

Here’s the kicker from that research: the more difficult the decision, the more likely we are not to act.

For a high performer, the decision to adopt AI is genuinely difficult. They have to weigh uncertain future benefits against certain current competence. And the research on status quo bias shows that people stick with their current retirement mutual funds even when those funds are underperforming. If people won’t switch investment vehicles when actual money is on the table, why would we expect your operations manager to eagerly abandon a workflow that’s been delivering results for a decade?

The endowment effect: “my way” feels more valuable because it’s mine

The endowment effect explains another layer of this resistance. Kahneman, Knetsch and Thaler showed in their 1990 research that people who own something value it at roughly twice what non-owners would pay for the same thing.

Your best performers don’t just use their current methods. They own them. They built them, refined them, invested in them. Their processes and expertise feel more valuable to them than they would to an outside observer, precisely because they’re theirs. An AI tool that promises to “enhance” their workflow is, from their perspective, asking them to trade something they value highly for something they haven’t yet experienced.

Fear of the unknown: the fundamental fear underneath everything

Underneath all of these biases sits something even more basic. Fear of the unknown is purported to be one of the fundamental fears that humans, and many other animals, share. When primed to feel uncertainty, our brains activate the amygdala, the part that controls our fear response.

Some researchers suggest this fear is so powerful that it could be the foundation of most other fears.

AI introduces enormous uncertainty. Not just “will this tool work?” but “will my expertise still matter?” and “will my role change in ways I can’t predict?” Your best performers have more invested in the current state of things, so the uncertainty feels proportionally larger for them.

The corporate antibody problem

When I work with organisations on AI adoption, I consistently see that the people resisting most are not the ones who can’t do the work. They’re often the ones who are best at it.

I’ve identified eight distinct types of corporate antibodies that resist innovation within organisations. Several of them map directly onto the behaviour you see from top performers facing AI:

  • The Status Quo Supporter values what’s currently working and fears how their job will change. Their concern is entirely legitimate: “Things work fine the way they are now.”
  • The Frustrated Elder has seen previous technology initiatives fail and knows their reputation is on the line if they back this one.
  • The Analytic Paralytic wants proof that AI will deliver before committing, because they know that most decision makers react negatively to highly creative or novel proposals.

None of these reactions are irrational. They’re built on millions of years of biological evolution to be self-protective. And data supports their caution: a Kyndryl 2024 survey found that 45% of CEOs report employee resistance or hostility to AI. A Leadership IQ study found 55% of employees are indifferent, resistant or actively sabotaging AI efforts. This isn’t a fringe reaction. It’s the majority.

The “ladder of no” makes it worse

There’s an organisational structure problem amplifying this individual psychology. I call it the Ladder of No.

In any hierarchical organisation, an employee with an idea for using AI needs to climb through multiple layers of management to get approval. Each rung on the ladder is someone who might say no. And you only need one “no” to kill the initiative.

Your best performers are often in the middle of this ladder. They’re senior enough to have influence, experienced enough to be listened to, and psychologically motivated (for all the rational reasons we’ve discussed) to resist. A single senior performer who says “I’ve been doing this for 15 years, and I don’t see how AI improves on what I deliver” can block adoption for an entire department.

How a CEO should think about this differently

Here’s where I see leadership teams make the biggest mistake. They treat resistance from top performers the same way they’d treat resistance from underperformers: as a training problem, a communication problem or a “mindset” issue.

It’s none of those things. And framing it that way actually increases the resistance, because it signals to your best people that you don’t understand or respect what they’ve built.

As I’ve written about the difference between how CEOs and department heads think about failure, there’s a fundamental gap in perspective. A CEO can think about AI adoption at the portfolio level: even if most experiments fail, one big success will pay for everything else. But each individual performer is thinking about their own personal success or failure. And a 90% failure rate at the portfolio level means each individual faces a 90% chance that their specific effort won’t work out.

Here’s what to do instead:

  1. Lower the perceived risk, not the intelligence of the resisters
    Stop framing AI adoption as “the future you need to get on board with” and start framing it as a tool that enhances what already works. The key insight from status quo bias research is that people become more accepting of change when exposed to it gradually. The mere exposure effect shows that the more often someone encounters a new idea, the more positively they see it.
  2. Start with quick wins that respect existing expertise
    In my consulting work, I’ve found that the fastest way to overcome resistance is to show, not tell. A quick win can be as simple as showing a top salesperson how AI can draft the first version of a proposal they’d normally spend two hours on. They still review it, they still own it, they still add their expertise. But the drudgery is reduced. This counteracts loss aversion because it demonstrates gain without requiring loss.
  3. Involve resisters in the process early
    For every type of corporate antibody, one of the most effective countermeasures is bringing them into the innovation process early. When your best performers help shape how AI gets implemented, they trigger the endowment effect in your favour. They start to own the change, and things you own feel more valuable.
  4. Reduce the rungs on the ladder of no
    If AI adoption has to pass through six layers of approval, it will die. Set up what I’d call an AI exploration pool: dedicated time and resources for people to experiment with AI tools in their own work, with minimal bureaucratic overhead. Your best performers don’t need permission to be excellent. Don’t make them ask permission to experiment with new tools either.
  5. Get a clear picture of where you actually stand
    Before trying to push adoption, understand the real landscape. Who on your team is enthusiastic? Who is sceptical? What are their specific concerns? An AI opportunities audit can surface these dynamics systematically, giving you a baseline that respects the reality of where your organisation is rather than where you wish it was.

The counterintuitive truth

Here’s what most AI consultants won’t tell you: your experienced people who resist AI are actually your biggest asset in adoption, not your biggest obstacle. They know where the real problems are. They understand which tasks are genuinely repetitive versus which ones require the nuanced judgment they’ve developed over years. They can tell you which AI outputs are good enough and which are dangerously wrong.

The key is treating their resistance as valuable signal rather than noise to overcome. When they say “AI can’t do what I do” they’re often right about the most important parts of their work. The opportunity isn’t to replace what they do. It’s to remove the friction around what they do so they can do more of it.

That requires a genuine AI strategy and governance approach that accounts for the human psychology at play, not just the technology capabilities on offer.

Your best performers built their careers by being excellent at what they do. Resistance to anything that threatens that excellence isn’t a flaw. It’s a feature. The question isn’t how to break through their resistance. It’s how to make adoption feel like an extension of the excellence they’ve already proven, rather than a replacement for it.

Frequently asked questions

Why do my most experienced employees resist AI more than junior staff?

Junior employees have less invested in current methods. They haven’t spent years refining a system that works, so adopting AI feels like adding a skill rather than risking one. Senior employees face stronger loss aversion because they have more to lose: reputation, mastery, proven results. Research from Tversky and Kahneman shows losing something feels roughly twice as bad as gaining something of equal value, and your best performers have accumulated far more to lose.

Is AI resistance a sign of low digital literacy?

Usually not, especially with top performers. Resistance from high performers is more often driven by the endowment effect and status quo bias than by lack of technical understanding. Many resistant senior staff understand AI perfectly well. Their objection is that the uncertain benefits don’t justify the disruption to methods that are demonstrably producing results. That’s not ignorance. It’s risk assessment.

Should I mandate AI adoption across the organisation?

Mandates tend to trigger exactly the biases that cause resistance in the first place. Forcing change amplifies fear of the unknown and activates corporate antibody behaviours. A more effective approach is to create space for experimentation, demonstrate quick wins and involve sceptics in shaping implementation. People adopt change more readily when they feel ownership over it rather than having it imposed on them.

How do I measure whether AI adoption resistance is hurting performance?

Look at the gap between teams that have adopted AI tools and those that haven’t, but be honest about what you’re measuring. If your resistant top performer is still outperforming colleagues who’ve adopted AI, the urgency for their personal adoption is genuinely lower. Focus instead on identifying repetitive, time-consuming tasks where AI can free up their capacity for the high-value work only they can do. That reframes adoption from threat to opportunity.

What’s the single most effective thing a CEO can do to accelerate AI adoption among reluctant top performers?

Remove the career risk. If your organisation punishes failed experiments harder than it rewards successful ones, no amount of training or communication will overcome loss aversion. Create explicit protection for people who try AI tools and find they don’t work for their use case. When the downside of experimenting is zero, the rational calculus shifts, and your best performers will start exploring on their own terms.