AI Roadmap
& Leadership

Situation

A global manufacturer with sites in four countries had something running with AI at every location. Quality inspection with image recognition in Germany, predictive maintenance in the Netherlands, a quoting assistant in Poland. All seven pilots worked technically. None had scaled.

Challenge

There was no owner. The CIO had inherited the portfolio, with no time and no AI background. Without someone guarding the agenda, every initiative faded the moment the local enthusiast got pulled onto something else. Hiring a full-time Head of AI was not realistic while the scope of that role was still unclear.

Approach

I started as fractional AI lead for three days a month. The first month went into an honest inventory: what is actually running, what does it cost, and who uses it. Two of the seven pilots turned out to have already quietly stopped. In months two and three we set up governance: one steering group, one decision framework, one set of metrics that all four countries reported against.

Decision

Three pilots got the green light to scale, two were deliberately ended, and two were parked until the data foundation was in order. The decision to stop cost the most discussion and freed up the most capacity.

Result

The roadmap was approved with a 25% productivity target on the processes involved. Three applications now run in production across multiple sites. After nine months an internally promoted lead took over the role; I scaled back to one day a month as a sounding board.