Almost every PE firm now claims to use AI, but Schroders Capital CIO Nils Rode thinks some are asking the wrong question about it.

The firm, which has $112 billion AUM and is the private markets and alternative investment division of asset manager Schroders, published a white paper in May arguing that AI’s highest-value contribution in PE will not come from cost savings or leaner teams, but from shifting the return distribution—finding more winners and avoiding more write-offs.

The arithmetic behind that argument is stark. According to the paper, which draws on a global sample of buyout investments studied over more than three decades, roughly one in four buyout deals delivers a gross IRR above 50% over a typical four-year hold, while one in 10 returns nothing at all. In an asset class driven by outliers, a single missed winner or an avoidable impairment can have a material impact on the fund.

Schroders Capital has developed three proprietary AI tools embedded in its investment process, including a GP screening tool called the Long/Short List, which surfaces managers matching patterns associated with historical outperformance across roughly 1,000 US and European buyout managers, alongside two AI agents, GAiiA and Vicky.

PitchBook spoke with Rode about how those tools work, where the industry stands, and why he believes firms focused on cost reduction are missing the bigger picture.

This interview has been edited for length and clarity.

PitchBook: There is a lot of industry talk about AI driving efficiency. What do you think about AI’s role in private equity?

Nils Rode, CIO of Schroders Capital

Courtesy of Schroders Capital

Nils Rode: We have a strong view there. We believe that these tools will much better use the available data, and they will lead to an explosion of analysis—good analysis—100 times more, and at the same time 100 times faster. So things that might have taken weeks you can do now within hours.

It will be a bit like PowerPoint. When PowerPoint launched, it did not mean people did the same presentations at 10% of the cost, but they did hundreds of thousands of times more presentations.

And so we believe the same will happen in our industry. It will lead to better investments.

PitchBook: How specifically does AI help make better investments?

Rode: If you just get one more great investment or you avoid one bad investment in a fund, that changes the performance of the fund. That has a huge impact on the performance for LPs and on the carried interest, and that impact is much higher than any cost saving that anybody could have.

Sign up for The Europe Pitch

Get our daily digest of private capital markets in the EMEA region.

Subscribe

PitchBook: How does Schroders use the AI agents to enhance performance?

Rode: We have two main AI agents. The first one is called GAiiA, that is a Generative AI Investment Analyst. That one is helping with data room analysis and due diligence. And the other agent is called Vicky, the Virtual Investment Committee Agent. That one knows all past investments, all the investment memos, but also the outcomes, which were the good ones and which were the bad ones, and then asks challenger questions to draft investment memos.

The way it goes is GAiiA could do a first draft of an investment memo, then Vicky can challenge it, based on investments we’ve done in the past, these are risks that might have been overlooked, or that’s worth digging deeper. Then [it] gives that feedback back to GAiiA. GAiiA does a new version of the investment memo.

Vicky, the challenger tool, is for avoiding bad investments because it tries to identify all the inconsistencies, risks and issues. It slows down the investment decision, because it raises more questions, and this can result in better analysis and it could also lead to investments not being done, because there are more doubts.

In between, there’s always a human who can oversee.

PitchBook: Why is human oversight so important?

Rode: Our strong view is that these systems will never have full autonomy, so never voting rights, never do any decision in the process, but that it will always be about assisting the humans. The people dimension in our business is very important. It’s about trust networks and relationships. Ultimately, companies are managed by people, and you need to read the management and the culture. There’s always a lot of information that is not in the data room.

Humans still have more context for an investment than the tool. Where the tool is getting better than humans is that I’m on the investment committee, and I’m with the firm for more than 20 years, and I should know all 300-plus direct investments that we have done, including all pages of each memo, and all outcomes of all 300, but of course no human can, but the system can. There are some areas where the systems can be better than humans. That’s where they need to focus.

PitchBook: What does it actually take to master AI in PE?

Rode: Mastering AI means not using a chatbot; it means applying your own philosophy, your own criteria, your own process, your own thinking to investments with the help of AI, and by using proprietary data. Not so many firms will be successful with that. But for those who do, it will lead to better investments.

This article originally appeared on PitchBook News