There's a pattern that plays out in organisations attempting AI adoption. A senior leader returns from a conference energised. A working group is formed. A tool is procured. A pilot is announced. And then, quietly, momentum stalls. Adoption is low. The tool sits underused. The narrative becomes "AI didn't work for us."

The problem rarely starts with the technology. It starts much earlier.

The Technology-First Trap

Most AI projects are framed as technology projects. They're led by IT or data teams, evaluated on technical criteria, and measured by deployment rather than adoption. The assumption is that if you build it — or buy it — people will use it.

They won't. Not reliably. Not at the scale that delivers real value.

This is because the barrier to AI adoption isn't access to tools. It's confidence, curiosity, and cultural readiness. People who don't understand what AI can and can't do will find reasons not to use it. People who feel threatened by it will quietly resist. People who weren't involved in choosing it will feel no ownership over making it work.

Technology arrives in the organisation. But the organisation isn't ready for it.

Why Everyone Needs to Be Involved

The AI projects that succeed share a common characteristic: they aren't led by technology teams alone. They're led by the people who understand the actual problems — editorial teams who know where decisions are hard, marketing teams who know where time gets wasted, operations teams who know where things get stuck.

When AI adoption is owned by IT, it tends to stay in IT. When it's owned by the whole organisation, it finds its way into the places where it actually makes a difference.

This isn't just about buy-in. It's about problem identification. The most valuable AI applications in publishing aren't the ones that technologists design in isolation — they're the ones that emerge when people across the business are given the confidence and the space to ask "could AI help with this?"

That question can only be asked by people who feel capable of asking it. Which brings us back to the starting point.

Confidence Before Tools

The organisations that get the most from AI investment are the ones that build confidence before they deploy tools. That means training that is practical and accessible rather than technical and abstract. It means creating space for experimentation without pressure to deliver immediate results. It means treating early use cases — however small — as proof points that build momentum rather than pilots that must be defended.

It also means being honest about concerns. In publishing, AI raises genuine questions about creativity, copyright, and jobs. Organisations that ignore those concerns or dismiss them lose trust early. Organisations that address them directly create the psychological safety that makes genuine experimentation possible.

What Good Looks Like

The organisations that navigate this well tend to follow a similar pattern. They start with awareness — helping people understand what AI is and what it isn't. They move to confidence — practical skills, hands-on experimentation, real examples from within the industry. Then they open up the conversation about use cases — not a technology team deciding where AI should go, but the whole organisation identifying where it could help.

From that conversation, real priorities emerge. Ideas get evaluated not just for technical feasibility but for practical value and economic benefit. The ones worth building get built. The ones that aren't get set aside. And the organisation develops a shared sense of what AI is for — which is the foundation everything else rests on.

The tools matter. But they're the last piece, not the first.