A different race 

AI changes not only how work is done, but where value remains. A language model, underneath everything, is trained to recognise patterns and predict what comes next. That tells you what AI will commoditise first: not analysis in general, but the conventional answer to the standard question. The median.

Producing the median answer has never set anyone apart. Until now, it has still been enough. Competence was scarce, took years to acquire and could be sold at a decent price for a working lifetime. Almost every professional career has been built on that fact. What is changing is that AI can now produce that answer to a high standard, across a breadth no individual can match, almost instantly and at negligible cost. Competing with it on those terms is not a contest humans can win.

The dominant objective in investment management is to beat the market. Nearly every asset manager is measured against it, sells against it and organises itself around it. The industry offers thousands of different funds, but most are ultimately variations on the same promise.

AI makes that promise harder to keep. As firms gain access to increasingly similar analytical capability and apply it to the same companies, information and benchmarks, the differences on which outperformance depends become harder to sustain. AI cannot make everyone a winner. It will instead compete away much of the advantage on which the race was built.

But there is a question that comes before all of that. Was beating the market ever what every client actually wanted?

Not all of them. Many care less about beating a benchmark than knowing their capital can provide an income or meet a future goal. Yet an industry organised around outperformance offers those investors surprisingly few alternatives.

Today's market makes the problem especially clear. Ten companies now account for almost 40% of the S&P 500, a level of concentration not seen in decades. Many sit at the centre of the same underlying bet: who wins the AI race. The industry may offer thousands of products, but an extraordinary amount of client capital is increasingly riding on the same outcome. It is the same risk packaged in different ways. For investors who value greater certainty, that leaves an important need unmet.

Meeting that need means running a different race. Not a more conservative version of the same one, but one with a different objective. Most of the industry defines success by doing better than someone else. We define it by whether our clients achieve what they set out to achieve. That requires good returns, but also outcomes dependable enough to plan around. Our aim is to achieve both: strong returns within a narrower range of possible outcomes.

Our investment approach therefore focuses on high-quality businesses whose earnings and dividends are durable enough to be relied upon, and on recognising quickly when that changes.

This is where AI earns its place with us. It monitors every holding, applies the same standard to each and flags deterioration in hours rather than weeks. Narrowing the range is mostly a matter of not being surprised, precisely the kind of work at which a system that does not tire excels.

For all the legitimate concerns about AI, I remain optimistic about what it can enable. By making capabilities that were once scarce widely available, it gives people and businesses with something distinctive to contribute a better chance to compete with larger, more established institutions. It also makes diversity of thought more valuable, not less. When conventional competence is abundant, value comes from what you add, not where you rank. Marriott has never been conventional. We were built to meet a need much of the industry had overlooked. That gives me confidence that we will continue to play a meaningful role for our clients in the age of AI.

Thank you to everyone who has taken the time to read this series. I hope it has offered something useful, and perhaps a different way of thinking about what AI may make possible.