How we approached it

At Marriott, we have spent the past twelve months working through that question. What emerged was not a technology project. It was a rethinking of how human judgement and AI capability work together.

The starting point was a conviction we have always held: the best decisions come from different perspectives, tested through honest debate. AI did not change that conviction – it extended it. People bring experience, judgement, and the ability to weigh things that cannot be measured. AI brings analytical depth, consistency, and the ability to monitor change at a speed no team could match. Our approach centred on getting the best out of both.

Five principles emerged from doing this in practice.

Principle 1: AI provides the base. Humans elevate it.

AI handles the repetitive work: processing information, applying consistent frameworks, watching for change. Our analysts focus on interpretation, context, and the decisions that require experience and intuition. This is not the automation of judgement; it is the elevation of it.

Principle 2: How we engage shapes what we get back.

We engage with AI deliberately, with curiosity and openness, rather than as a tool to be commanded. A thoughtful question produces a richer answer than an instruction. This is how we work with each other. It is how we work with AI. For clients, this leads to more thoughtful insights, not just faster answers, improving the quality of the investment decisions we make on their behalf.

Principle 3: Every output is controlled and traceable.

Our AI operates within tightly defined boundaries. Each analysis uses only the data it is given, follows fixed rules, and produces a clear record of how it reached its conclusions. If something informed a conclusion, you can see exactly what it was and the reasoning behind it.

Principle 4: Attention goes where it matters most.

Our system identifies where judgement most improves outcomes and directs analyst time there: ambiguous situations, signals that require context, shifts that numbers alone cannot explain. We want our analysts spending their time on the five to ten percent of work AI cannot handle. That is what getting the best out of both looks like in practice – more time spent on the variables that matter most.

Principle 5: No AI output reaches a decision without human approval.

Our Investment Committee has full authority over every significant decision, through the same culture of collective review that has governed our process for more than twenty years. Nothing is automated. Every override is documented. Accountability sits with the people who carry it. Clients retain the assurance that every investment decision remains firmly in the hands of experienced professionals.

What this means for our clients

For our investors, this means an investment process better equipped to protect capital, detect risk earlier, and deliver more consistent and predictable outcomes, with capability that continuously strengthens as AI capability advances.

Firms that are not willing to reimagine their approach will remain constrained by existing processes, limiting what technology can meaningfully improve. The technology is not the bottleneck. The willingness to rethink is. Firms that recognise this and build accordingly will operate with the breadth and consistency of a global research platform, without the headcount and structural rigidity. When the capability exists today, waiting has a cost. That cost compounds.