By Duggan Matthews, Chief Investment Officer,
Marriott Investment Managers
This is the fourth article in a series on what AI means for investment management. The earlier pieces argued that the advantage lies in how you structure AI rather than in having it, and that Investment Management has been slow to act on this. This piece is about what comes next.
This article comes to you from Cambridge, Massachusetts, where I had the privilege of attending the executive programme on AI at MIT. It was an extraordinary two weeks spent with sixty business leaders from over twenty countries, working across a wide range of industries, taught by some of the brightest minds on the subject in the world.
The mood in the Sloan rooms was unlike anything I have experienced. Sixty people in mid-career, sitting with the realisation that the defining technology of our lifetimes is moving faster than most of us have absorbed.
Amara's Law states that we tend to overestimate the impact of new technology in the short term and underestimate it in the long term. Given the current level of hype around AI, the idea that we might still be underestimating the long-term implications is hard to take in. But two weeks at MIT made it clear that we probably are.
I will not try to recite everything I learned. I want to focus on one thing.
AI has already matched or exceeded human performance across a wide range of expert tasks, in some cases by a significant margin. That gap is going to widen. Hundreds of billions of dollars are being invested, and the scaling laws that have driven recent progress still hold. Whatever you think AI can do today, it will do more tomorrow.
This leaves an obvious question. What is left for us?
The instinct is to answer by listing the things AI cannot do. But that is the wrong approach. The capabilities of AI will keep changing, and any answer based on what it cannot do today will be out of date soon enough. The better question is different. What should remain human, and why?
In the investment management industry, one of the obvious answers is engaging with management teams. Spending real time with the people running a business builds an understanding that no transcript or filing can replicate. You get a read on how they think, how they respond to pressure, and what kind of people they are, not just what they say.
This is a small example of a much larger pattern. The most valuable information in investing, and I would argue in most serious work, is rarely the information that already exists. It is the information that emerges from interaction. From two people with different mental models pressing on the same problem. From the moment one person sees what another has missed. From the debate that forces a sharper answer than either started with.
Recent research from Procter & Gamble and Harvard1, which was extensively discussed and debated at the MIT programme, makes the point directly. They ran a large experiment comparing individuals and teams, with and without AI. Individuals using AI matched the average quality of human teams without AI, which is a striking result on its own. But the more important finding was about exceptional output. Teams using AI were roughly three times more likely than the control group to produce top-decile solutions. Individuals using AI were not. AI alone raises the floor. Teams with AI raise the ceiling.
The implication for investment management is significant. The traditional model of a single portfolio manager, however experienced, making all the calls is unlikely to survive this transition. At Marriott, for over twenty years we have run all our portfolios through an investment committee rather than a lead portfolio manager. The committee seeks consensus on every decision. That choice was made for reasons that had nothing to do with AI, but the logic that supported it then is the logic AI is now making unavoidable.
AI is already smarter than any individual portfolio manager. It is not smarter than a team of people with genuinely different perspectives, because a team is not a single intelligence but the friction between several. What remains, after the analytical edge has been commoditised, is the value of that friction. The right answer matters less than the right question, and knowing what to ask is itself a team activity.
This brings me to the part I want to spend the most time on, because I think it matters more than the rest. Diversity.
In a world where the analytical floor is commoditised, the only way to reach the ceiling is through genuine variety of perspective. The same Procter & Gamble and Harvard research showed something quietly important. Without AI, teams produced outputs that clustered around whichever team member dominated the conversation. With AI, that dominance effect faded, and the variety of perspectives in the room actually surfaced in the final work. AI did not reduce diversity. It made it more productive.
That changes the argument. Diversity is what will separate the firms producing top-decile work from the ones producing the AI-floor average that anyone can replicate. It moves from being a virtue to a mechanism for better decision making and outcomes.
Which means the question stops being how many people you have, and starts being how different they are from each other and from the AI they are working with. You only have to look at the Springboks to see what this can look like. Rassie Erasmus did not build the most diverse squad in world rugby because it was admirable. He built it because he believed it was the only way to win at the level required, and he was right. The variety of backgrounds, languages, and styles of play was not something the team overcame to succeed. It was the reason they succeeded.
The implication for how we develop and shape the investment team at Marriott in the years ahead is clear. As AI commoditises analytical capability, our advantage will come increasingly from the variety of thinking in the room and the culture that lets that variety actually function.
This is what I come back from MIT thinking about. AI, no matter how brilliant, will ultimately be a commodity, because everyone will have it. In time, the architectures we build around it will commoditise too. The advantage cannot live in either. It will live where it has always lived, in the people. Not in the same way as before, and not doing the same things as before, but in the human capacity for judgement, interaction, and genuine variety of thought, sharpened by tools and channelled through the systems we design around them.
The technology will keep moving, and that part will take care of itself. What will separate the firms that thrive from the ones that drift is something much older than AI. The willingness to put genuinely different minds in the same room, to give them the tools to do their best work, and to build the kind of culture where the best idea wins regardless of who said it. The same logic applies wherever judgement matters, which is to say, almost everywhere.
1 Recent research from Harvard Business School, conducted in collaboration with Procter & Gamble and Wharton, examined 776 P&G professionals working on real product innovation challenges, either individually or in teams, with or without AI support.
Local
Offshore