I set the strategy, own the governance, and build the teams that deliver.

Most organisations have more AI ideas than capacity. My job is to decide which ones are worth backing, in what order, and why.

I sit with the people closest to the work, find where cost, effort and risk really concentrate, and test every idea against three questions: is the data there, is the value real, and can we live with how it might fail?

The result is a sequenced, costed roadmap that executives can fund with confidence, and that the team understands well enough to deliver.

Governance only works when it is built into delivery, not bolted on at the end. I put the guardrails where the work happens: validation before a model ships, monitoring once it is live, and clear rules for when a person must stay in the loop.

When a board, auditor or regulator asks how a decision was made, the answer is ready: the lineage, the testing and the sign-off.

Done well, governance is what lets teams move faster, because everyone already trusts the system to fail safely.

AI efforts rarely fail on the model. They fail on the team around it: unclear ownership, no shared standards, talented people pulling in different directions.

I set the structure that fixes that: clear roles, a real definition of done, code and model review, and a steady planning rhythm. I hire for the gaps that matter and coach leads to hold the standard themselves.

What I leave behind is a high-trust team that keeps delivering, whether I am in the room or not.

A model in a notebook is not an outcome. I keep every piece of work tied to a business result, from the first prioritisation call to the release that customers or staff actually use.

That means shipping the smallest useful thing early, measuring whether it moves the number that matters, and doubling down only on what works.

My track record runs from predictive models that guided investment and policy to AI products in real use.