Asia changed my perspective
What Six Months in Asia Taught Me About the Limits of My Own Experience I stepped off a plane in Hong Kong in early 2025, carrying twenty years of assumptions about how regulated industries work, how risk is managed, how governance frameworks are built, and, if I am honest, about where the serious thinking in financial services actually happens. I had built a career in London, operated across EMEA, navigated Basel frameworks, sat in front of regulators in three time zones. I thought I had a reasonably accurate map of the global landscape. By the time I landed back at Heathrow six months later, that map was in pieces on the floor. This is not a story about being humbled by the exotic East. That framing is its own form of condescension, and it is not what happened. What happened was more specific, and more uncomfortable: I discovered that some of the hardest problems I had spent years advising organisations to manage had already been solved – operationally, not theoretically – in places I had been too comfortable to spend serious time in. The Moment I Stopped Being Certain The jolt came in Singapore, about six weeks into the trip. I was in a working session with a team running AI governance infrastructure for a major financial institution. Not a pilot, not a roadmap: live infrastructure, being stress-tested against real regulatory requirements in real time. The team lead was explaining their model validation approach – the controls architecture, the feedback loops, the way they had structured human oversight into the decision chain – and I was taking notes like a junior analyst. She was thirty-two, maybe thirty-three. In London, the person with that responsibility is usually someone with grey hair, institutional scar tissue from a crisis or two, and a carefully curated network of regulators they can call. Experience is used as a credential. Here, the credential was demonstrated capability. The question the organisation had asked was not “who has done this before?” It was “who can actually do this now?” I have spent two decades in rooms where seniority determines credibility. That session in Singapore was the first time in a long while that I felt the distance between my assumptions and reality as a physical thing. Three Things That Refused to Fit My Existing Model I started in Hong Kong, because it was where the dissonance started. The regulatory environment there operates at a pace I was not prepared for. In London, the gap between regulatory intent and enforcement action can stretch across years – consultations, industry responses, phased implementation, guidance notes on the guidance notes. In Hong Kong, that gap is measured in weeks. The senior executives I met were doing something I found genuinely rare: they were holding two entirely different regulatory worldviews simultaneously – mainland and international – without it visibly destabilising their decision-making. That is not a skill most frameworks teach. It is the product of operating under sustained complexity for long enough that ambiguity becomes a normal working condition rather than a problem to be resolved before you proceed. Singapore was infrastructure where I expected aspiration. AI governance in that market is not a strategy document. It is funded, staffed, and running. The Model Risk Management frameworks had been adapted specifically for generative AI contexts – not retrofitted from credit risk models from 2009, which is what I see most often in European institutions. The regulatory bodies had developed technical capacity in parallel with the private sector, not after it. That sequencing matters more than most governance discussions acknowledge. Tokyo took longer to read, but it had the most to say about execution. The pace was slower. Consensus takes the time it takes, and there is no shortcut that does not eventually cost you. But the data governance practices I observed had a quality I had genuinely not expected: they were operational, not decorative. There was no gap between policy and practice, no shelf full of frameworks that the business quietly ignores. The controls were embedded in how work actually happened. In my experience across UK and European financial institutions, that gap – between the governance document and the governed reality – is one of the most persistent and expensive problems in the industry. In the organisations I visited in Tokyo, it had been closed. They were not debating data quality. They had built systems that made low-quality data structurally difficult to introduce. The Thought That Arrived Somewhere Over the Gulf On the third return leg – somewhere between Dubai and London, around 2am – something settled. Asia is not catching up to Western regulatory and governance thinking. That framing assumes the West defined the destination and everyone else is navigating toward it. What I had actually witnessed was a set of jurisdictions solving problems that the West has not yet named clearly enough to begin solving. The West is good at frameworks. I write them, export them, consult on them, and convene conferences about them. The frameworks are often genuinely rigorous. But a framework without the infrastructure to run it is a very expensive piece of intellectual comfort. What I saw in those six months was the infrastructure – built with urgency, staffed for capability rather than seniority, and designed to adapt rather than to endure. What This Means for Anyone Leading in Risk, Compliance, or AI If my organisation is building AI governance, updating its model risk framework, or trying to understand how regulatory expectations are shifting globally – the most useful thing I can do is not commission another benchmarking report from a global consultancy. Those reports will tell me what was true eighteen months ago, filtered through a lens that probably originated in Western financial centres. I will go. I will sit in the actual rooms. My education begins the moment I clear customs and starts to compound when I realise how many of my own assumptions were doing work I never asked them to do. The competence