Almost every conversation these days includes some version of the same question: "How are you using AI in your investment process?" It's a fair question, and it deserves a better answer than the vague answers most of the industry offers. As a firm that has spent the past couple of years actually working with these tools, the honest answer is that the reality is more mundane than the marketing implies, and more useful for exactly that reason.
Let us start with what AI is not doing for us.
It is not picking stocks. It is not sizing positions. It is not deciding how much risk to carry into a binary event, or whether a sum-of-the-parts discount reflects genuine complexity or simple market neglect, or when a cheap stock is cheap for a reason. Those aren't clerical tasks that happen to sit at the end of the process, they are the process. Any firm that outsources them to a model hasn't removed risk; it has moved the risk somewhere less visible and harder to challenge.
What has genuinely changed is the cost of a first pass.
Reading a 40-page initiation report, a foundry capex disclosure, or three years of audited financials used to consume pure analyst hours. Now the first read takes minutes. This doesn't eliminate careful reading, rather it changes where scarce attention gets spent. If a tool can get you to the right sections faster, flag inconsistencies between a new disclosure and an old one, and frame the obvious questions before the deep work starts, the analyst spends more time on the second and third reads, which is where judgment actually lives. Edge was never built by knowing a signal exists; it comes from working out whether the signal matters, what the market has already priced, and what could shift the distribution of outcomes. Nothing about that has changed.
Monitoring works the same way. A diversified global book generates a constant stream of small signals, including filing changes, a shift in options positioning, a regulatory update, a local news item that never reaches the global feed. AI is genuinely good at flagging that something changed. It is much less reliable at telling you whether the change is material, and confusing those two capabilities is how mistakes get made faster rather than avoided.
There's a less comfortable point that deserves airing: AI makes mediocre work look polished. A clean summary can create the illusion of understanding. A confidently written paragraph can conceal weak sourcing, stale data, or one bad assumption doing all the load-bearing work. In this business, being approximately right for the wrong reason is often worse than being openly uncertain, because the wrong reason eventually catches up with you at size. So, the process has to treat AI output as a draft, never as evidence. The burden of verification doesn't fall away because the paragraph reads well.
Where the tools have surprised us on the upside is as a forcing function. Asking a model to argue against a thesis, to identify the single assumption the whole idea rests on, or to list what's missing from an analysis is cheap, fast, and occasionally uncomfortable in a productive way. It won't find what a good analyst misses every time. But it raises the floor on preparation, and preparation is where most avoidable errors originate.
One structural risk is worth naming. If most market participants end up running similar models on similar data, correlation risk can rise even while individual decision quality appears to improve. Regulators have started thinking out loud about what happens when automated systems all reach for the same trade at once. That isn't a reason to avoid the technology, rather it's a reason to be deliberate about where it sits. In our view, AI belongs near the front of the funnel, where breadth, retrieval and speed matter. It doesn't belong at the point of final judgment, where context, incentives, liquidity, sizing and plain nerve matter most.
So, the candid version of "how we use AI" is this: it compresses research time, improves monitoring, gets us to the right questions faster, and makes the first draft of almost anything much cheaper. It does not create conviction, it does not know when the market is missing something, and it does not carry the consequences of being wrong.
The firms that get hurt by this technology won't necessarily be the ones that ignored it. More likely, they'll be the ones that adopted it enthusiastically but carelessly by letting it quietly absorb decisions nobody explicitly meant to hand over. The opportunity, properly understood, runs the other way: use the machines to protect the scarcest resources in the process, the time, attention and judgment of the people making the decisions.