What 6,000 Private Equity Professionals Taught Me About AI in Berlin

I just returned from Berlin, where I had the privilege of serving as Chair of the AI and Technology Innovation in Private Equity Summit at SuperReturn International 2026. SuperReturn is the world’s largest gathering of private equity professionals, and this year drew more than 6,000 attendees from over 60 countries to the InterContinental Hotel on the Budapester Strasse. It is, without question, one of the most concentrated rooms of financial intelligence on the planet.

I moderated and participated in several panel discussions over the course of the day, alongside some genuinely sharp minds from firms like Google, Amazon Web Services, IBM, EY-Parthenon, Bain Capital, KKR, and a handful of others. The topic threading through every session was the same one that threads through everything we do at Business Laboratory: how do you take the enormous promise of AI and turn it into something that actually moves the needle for a business?

In private equity, that question carries a particular urgency. And the answers I heard in Berlin confirmed several things I already believed, while genuinely surprising me on a few others.

The Gap Between Ambition and Execution is Enormous

The first panel I moderated was titled “Overcoming Execution Risk in Operational Initiatives,” and I will be honest with you: I could have given this talk myself without any of my excellent co-panelists, because we have lived this story in our own client work for years. But the fact that this topic was on the main agenda at SuperReturn, in front of the most sophisticated financial audience in the world, tells you something important. Even the very best-resourced firms are struggling to close the gap between what AI can theoretically do and what they have actually managed to make it do.

The reasons are familiar: lack of connection to the core value chain; poor implementation planning; business cases that look compelling on a slide deck but fall apart on contact with real operations. A failure to get genuine ownership behind an initiative from the people who will actually have to live with it. These are not technology problems. They are people, scope, methodology, and process problems, and no amount of additional GPU horsepower fixes them.

What struck me most in that panel was the consensus that the firms making real progress with AI are the ones that have slowed down enough to understand their operations before layering AI on top of them. The ones that are struggling are the ones that bought a tool and went looking for a problem to attach it to. That distinction is not subtle and it is not new, but it bears repeating because the temptation to do it backwards is apparently irresistible, even for sophisticated investors managing hundreds of billions of dollars.

Due Diligence is Getting a Stress Test

The panel on tech due diligence under compressed timelines was the one that generated the most animated discussion of the day. The problem is real and getting worse. Private equity firms are being asked to evaluate increasingly complex technology companies in deal timelines that have not expanded to match that complexity. The result is that critical questions about a target company’s technical debt, data infrastructure, and AI readiness either get answered superficially or not at all.

What I found encouraging was the emerging consensus that automation is not just a nice-to-have in the due diligence process but a genuine competitive differentiator. The firms that are building systematic, repeatable approaches to technical assessment, ones that use AI to do the initial heavy lifting and reserve human judgment for the genuinely hard calls, are getting better answers faster than the ones still relying on a two-week sprint of consultant interviews. The game is changing and the firms that recognize it are pulling ahead.

The “Last Mile” Problem is Real

One of the sharpest presentations of the day came from a speaker who was not on any of my panels: Charlie Pickell from Hebbia, who spoke about what he called the “last mile imperative.” His argument, delivered crisply and without any of the usual AI hype, was that the value of AI keeps getting lost somewhere between the individual analyst who understands the tool and the institution that needs to benefit from it. The individual gets smarter. The firm does not.

I have seen this exact dynamic in our own client work, and I think it is one of the most underappreciated problems in the whole AI conversation. Companies celebrate the fact that a few of their people are using AI productively without ever asking the harder question: what would it look like if the organization used AI productively? Those are very different things, and the gap between them is where most of the value is hiding.

What Emerging Markets Are Teaching the Rest of Us

Later in the afternoon I joined Noor Sweid, Founder and Managing Partner of Global Ventures, for a conversation drawn from a London Business School case study on value creation in emerging markets. This one surprised me the most, in the best possible way.

The conventional assumption is that AI innovation flows from the most resource-rich environments outward. What Noor’s work illustrates is nearly the opposite. Companies operating in capital-constrained, infrastructure-limited environments have been forced to build leaner, more precise AI applications than their Western counterparts, because they cannot afford the luxury of experimenting broadly and failing expensively. The result is a playbook built on surgical deployment (“surgical doses of AI” is an often-repeated phrase from me and my team), capital efficiency, and relentless focus on measurable outcomes. That, as it turns out, is exactly the playbook that works everywhere.

The best AI is not the most expensive AI. It is the AI that is placed in precisely the right spot within a system that is clearly understood, with a specific outcome in mind, by people who know the business cold. Often the best AI is that which is completely invisible to the human workers within the firm, working autonomously in the background.  Every emerging market innovator in that room already knew this. I am not sure the rest of the room has fully absorbed it yet.

What I Brought Home From Berlin

I have been making the case for years that AI is not a product you buy. It is a capability you build, carefully, in bespoke fashion, on a foundation of operational clarity. Berlin confirmed that this view is not a contrarian one anymore. It is becoming the consensus among the most analytically sophisticated investors in the world.

What I find genuinely exciting is that private equity, of all industries, has both the incentive and the analytical horsepower to get this right. When you are managing a portfolio of companies and your entire business model depends on making those companies measurably more valuable over a defined time horizon, “let’s run an AI pilot and see what happens” is not a strategy. Precision is a strategy. Measurement is a strategy. Understanding how your portfolio companies actually work, at a system level, before you decide where to apply AI is a strategy.

That is what the best firms in that room are figuring out. And frankly, it is the same thing we help our clients figure out, whether they are managing billions of dollars in private equity assets or running a mid-size manufacturer in the American heartland.

The problems are the same. The discipline required to solve them is the same. Berlin just reminded me of that at a very large scale, in a very nice hotel, with rather good German beer.

If you want to talk through what any of this means for your organization, you know where to find me.