AI adoption is a behaviour change problem.

I grew up between Copenhagen, the west coast of Ireland and Melbourne, and have lived in Amsterdam for thirteen years. Each move taught me the same thing: explore first, ask questions, and work out how a system runs before you try to change it.

That is what I brought to RELX when it was still unclear what AI would mean for marketing. I built the case, answered the objections before anyone raised them, and pitched a role that did not yet exist. They said yes, and that role became the engine for how a 200+ person team learned to work with AI instead of around it.

The hardest part was never the tools. It was getting AI written into people's actual objectives, so that it stopped being something they tried once and forgot. I did that one habit, one workflow and one development conversation at a time, until usage became simply how the team works.

1. i

   Workflow before habit, habit before objective

   Rebuild the work first, then make the new way the easy default, then write it into goals so it holds.

2. ii

   Specialist entry points for each team

   Different teams need different tools and different entry points. A single enterprise rollout is usually the slowest route.

3. iii

   Guardrails that create speed

   Clear governance gives people permission to experiment, which is what turns a pilot into an operating model.