Why "my team won't use it" is the most common AI failure - and how to fix it

I want to tell you about a conversation I had with a business owner a few months ago. He had spent around £15,000 on AI tools, training, and consultancy over the previous year. He had a custom-built workflow for his sales team, a document generation system for his operations team, and a content production setup for marketing. Everything worked. And almost nobody was using any of it.

"I don't understand it," he said. "We showed them how it works. We did the training. They just go back to doing it the old way." He was frustrated, and he had every right to be. But the problem was not his team. The problem was that adoption had been treated as an afterthought - something that would happen naturally once the tools were built. It never does.

"Adoption is not a training problem. It is a design problem. The system has to fit the way people actually work, not the way you wish they worked."

Why teams revert to old habits

There are three reasons I see teams failing to adopt AI tools, and they are almost always present in combination. Understanding all three is necessary before you can fix any of them.

The first is friction. If using the AI tool requires more steps than the old way - opening a different application, copying and pasting between systems, remembering a specific prompt format - people will not do it consistently. Humans are remarkably good at finding the path of least resistance, and if the old way is easier, the old way wins. The fix is to reduce the friction until using the AI tool is genuinely easier than not using it.

The second is trust. People do not use tools they do not trust. If the AI produces output that is sometimes wrong, sometimes off-brand, sometimes just not quite right, the team will stop relying on it - even if it is right 90% of the time. The fix is to build the trust through visible wins. Start with a use case where the AI is reliably good, let the team experience it working well, and expand from there.

The third is ownership. This is the one that most implementations miss entirely. When a tool is built by someone else and handed to a team with instructions on how to use it, the team's relationship with that tool is fundamentally passive. They are users of something that was done to them, not participants in something they helped create. The fix is to involve the team in the design process - not as a consultation exercise, but as genuine co-designers of how the tool fits their work.

What this looks like in practice

In the Foundation Programme, the first month is almost entirely about adoption. We build one workflow - the highest-leverage one we identified in the Discovery Workshop - and we spend the month getting the team to use it every day. Not adding more tools, not building more workflows. Just one thing, used consistently, until it becomes habit.

By the end of month one, the team has a working system they trust, a measurable result they can point to, and the confidence that comes from having done it once. Month two is significantly easier because of that foundation. Month three is easier still.

The businesses that get the most from AI are not the ones that implement the most tools. They are the ones that implement the right tools, in the right sequence, with the right level of attention to the human side of the change. Adoption is the work. Everything else is just setup.