People Don’t Resist AI Because They Don’t Understand It. They Resist It Because They Do.

The hardest part of any technology adoption has never been the technology. I learned this years ago leading a platform migration where the system went live on time, the infrastructure was solid, and the technical team was proud of what they’d built. The problem was the 200 people whose entire working day had been built around the old system. They weren’t resistant because they were difficult. They were resistant because change threatened something real, their competence, their routines, their sense of being good at their jobs. AI adoption is triggering the same dynamic at a much larger scale, and most organizations are responding the same way they always have: training sessions, comms plans, and a go-live date.

I’ve spent thousands of hours in change management, and the pattern plays out almost identically every time. The initial rollout generates visible excitement. A handful of early adopters sprint ahead and become champions. Leadership declares momentum. And then, quietly, the majority reverts. The licenses stay active. The dashboards show usage. But the actual workflow integration lags months behind what the numbers suggest. On one migration initiative I led, adoption looked healthy on every metric we were tracking, right up until we did a deeper audit and found that most people were copy-pasting AI outputs into the same manual processes they’d been using before. They were using the tool without changing how they worked. That’s not adoption. That’s compliance theatre.

What actually moves the needle is treating AI adoption the way you’d treat any serious organisational change; not as a technology project with a training module attached. That means doing the stakeholder mapping. It means identifying who feels threatened by what this tool implies about their value, and having those conversations directly rather than hoping an FAQ answers them. Some of the most important change management work I’ve done has been one-on-one conversations where I sat across from someone and said: here’s what’s not changing about your role. Here’s where this tool takes the pressure off, not where it takes over. Those conversations can’t be templated. They require a human being willing to be honest.

The most successful AI adoption I’ve been part of started not with a tool demo but with a question: what part of your week do you dread most? We let people answer that honestly; the reporting that ate up Friday afternoons, the status updates that required pulling data from four different systems, the first drafts nobody wanted to write. Then we showed them, specifically and practically, how AI could take that exact thing off their plate. The transformation in the room was immediate. People who’d been skeptical became the loudest advocates, not because leadership pushed them to be but because the value had landed personally.

The technology is the easy part. People adopt what makes their lives genuinely easier, resist what feels like it’s been done to them, and trust what they had a hand in shaping. The organisations that understand this will embed AI into how they actually work. The ones that treat it as a deployment milestone will be puzzling over adoption rates for years, wondering why the tools they paid for aren’t delivering the results they projected.