Results are not the same as progress.

The year that changed the question

I had been promoted after thirty days into a business I barely understood. The team knew the customers, the product and the history better than I did. Instead of earning their trust, I arrived with targets and a plan. Sales rose. People left.

The chairman’s verdict was blunt: I had delivered the result and damaged the organisation that produced it. He was right.

It would be convenient to say I understood that immediately. I did not. I left angry and defensive, and I never apologised to the Crosslink team. It took years, and smaller versions of the same mistake, to see that the failure was not despite the results. It was how I achieved them.

What that became

The Human Layer became the name for everything around a transformation that determines whether the result will last: trust, judgment, leadership, process, skills, data, governance and culture. The technology matters. So do the humans who have to make sense of it, challenge it and live with what follows.

AI does not repair an organisation. It accelerates what is already there. Clear leadership becomes more capable. Confused leadership becomes confused faster. A healthy culture learns. A fearful one gets better at hiding what is going wrong.

The framework did not emerge from a distance. It is a discipline I still practise. My first instinct is still to move too fast and default to outcomes. The work is noticing earlier, listening longer and building the relationships before they are needed.

The work now

That work now sits at AIR APAC, which publishes measures of AI readiness and of how AI systems represent places, together with the methods and limits behind them. The findings become keynotes, briefings and practical work with leadership teams. The question stays consistent: what should people decide, check and own before an AI system is allowed to act?

Earlier chapters included transformation work at HSBC, enterprise technology in the United States, and advisory work across Asia Pacific, the UK and Australia. Those experiences matter less as badges than as places where the consequences of decisions became visible.

This site publishes the evidence, the argument and the parts that did not work. Speaking engagements are open. Advisory stays inside relationships that already exist.

The line I hold

Anything that measures, scores, assesses or audits belongs to the institution. The method is published and the result is free to the organisation or place measured. It is never a personal product.

Speaking and advisory remain separate from that measurement work, including no engagement with an organisation during an open measurement period. The full rules are published at airapac.org/independence.