James Brocklebank has spent nearly three decades making billion-dollar bets on companies most investors would never touch — carve-outs from struggling conglomerates, businesses buried inside bureaucratic giants, deals closed at the height of a pandemic. He has seen complexity and leaned into it. But nothing, he says, compares to what’s coming with artificial intelligence.
“Excited and terrified in equal measure,” Brocklebank replied flatly when asked what he was most excited about in the world right now. Consider who’s saying this.
Brocklebank is co-chair of Advent International, a global private equity firm that manages roughly $100 billion and has backed some of the most consequential corporate transformations of the past two decades — from Worldpay to Thyssenkrupp Elevator to a portfolio of defense technology companies that include Shield AI and autonomous naval vessel startup Saronic.
Speaking on Goldman Sachs’s Exchanges: Great Investors podcast, he offered a candid window into how a top-tier investment firm is actually deploying AI — not as a marketing talking point, but as a tool already reshaping how deals get done.
The centerpiece: something Advent calls the IC Robot.
“This is an AI that has been trained on all of our investment committee papers over the last 13 years,” Brocklebank explained. “So, it’s ingested everything — all the deals that we’ve shown the committee. And interestingly, it has the perspective on what happened to the deals that we did, but also the deals that we didn’t do.”
Most institutional knowledge in private equity lives inside the heads of senior partners or buried in files no one revisits. Advent has effectively digitized its own intellectual history — its wins, its misses, its assumptions — and turned it into a live analytical layer that interrogates every new deal that comes before the committee.
When a new investment memo lands, the IC Robot reads it, cross-references the assumptions against 13 years of outcomes, and pushes back.
“It can basically say, ‘Oh, I see a margin assumption here that has not been achieved before in this particular type of company,’” Brocklebank said. “It’s not a voting member or anything like that. But it’s extremely powerful in terms of prompting discussion on particular topics.”
It is, in essence, a permanent devil’s advocate — one that never forgets a deal, never gets tired, and has no ego invested in making the numbers work.
‘Working for the AI’
Brocklebank was careful to frame the IC Robot as a prompt, not a decision-maker. But elsewhere in the conversation, he let slip a phrase that may prove more revealing.
“We’re almost moving to a landscape where we’re working for the AI rather than the AI working for us,” he said.
He was speaking with regard to data organization, as the AI will only be as effective as the structure that’s fed into it. Fortune contributor Jeffrey Sonnenfeld, of the Yale School of Management, has written at length about the key importance of data infrastructure.
“It’s also a cultural question,” Brocklebank said, because you need to break down silos in old ways of doing things and get the people on board with doing things in a different way.”
Advent hired a chief data science officer and a chief digital officer “before AI was a thing,” he noted. The firm built its own large language model — dubbed “Advent GPT” — years ago, before graduating to what he describes as a more sophisticated “intelligence engine” that draws on multiple LLMs simultaneously.
On the portfolio side, Advent has dedicated staff embedded in its companies specifically to drive AI transformation — not the “little AI projects bolted on” that Brocklebank dismisses, but a more foundational rethinking of how data is organized and how work gets done.
On hiring, Advent has already shifted what it looks for. “Rather than have a deal team and then a team of data scientists, we want the deal team to be already experienced in data analytics so that they’re natural with it, rather than a separate group,” Brocklebank said.
The human counterweight
What makes Brocklebank’s AI worldview unusual — at least among the finance elite publicly talking about the subject — is his insistence on keeping the human element at the center, precisely because he believes AI will eventually make it scarce.
Asked what advice he gives his sons as they consider careers in finance, he didn’t say learn to code or master financial modeling. He said: “Be human. In a world of AI, I think humanity is very important.”
The best defense against obsolescence, in his view, is emotional intelligence and relationship-building — the things an LLM cannot replicate. “The AI is like a superpower, a cloak of superpower” that will some day “give everyone the opportunity to do amazing things.” The difference, though, remains your core humanity and how well you know and use that, he added.
It’s a philosophy that fits neatly with how Brocklebank has always operated — leaning into complexity, trusting local knowledge, building long-term relationships in markets others find too opaque to navigate. He has applied that approach to Europe’s fragmented regulatory patchwork, to carve-outs in Germany and Spain, to a deal on Thyssenkrupp Elevator underwritten in March 2020 while the world was shutting down.
Now he’s applying the same logic to AI: embrace it earlier and more deeply than your competitors, while betting that the irreducible human advantage — judgment, trust, relationships — will determine who wins in the end.
“Excited and terrified,” he repeated. “In equal measure.”
For this story, Fortune journalists used generative AI as a research tool. An editor verified the accuracy of the information before publishing.
The post ‘Excited and terrified’: One of private equity’s top investors built an AI that knows every deal he’s ever done appeared first on Fortune.
