What professional investors can learn from Modernist artists on AI adoption

Most of the conversation about AI in private markets today is all about efficiency. From Private Equity and Credit to Family Office investing and M&A, leaders are focused on how fast their teams can source and diligence deals, make decisions and deploy capital. With the current AI landscape, the timeline for this entire process has been compressed and while the gains are real, the consequences are also important to consider.

Today’s leaders are asking whether AI improves investment decisions. However, the question they should be asking is what it does to judgment itself – how it's formed, how it gets passed to the next generation of investors and who's accountable for it.

Investment judgment is rarely taught in a classroom. Historically, it has been transmitted through junior work, and hours of it. Building a sourcing memo, assembling a diligence file, sitting in the room while senior partners weigh incomplete evidence and argue about what it means. An associate who spends three weeks inside a data room learns things an AI summary cannot carry over, such as which numbers move, which management claims survive contact with customers, and how much weight a partner puts on a founder's answer to an unexpected question. The apprenticeship model is how the industry produces decision-makers. It has stayed informal because it has never needed to be anything else

Former investor Alex Sen articulates this idea well. He spent nearly a decade investing at Blackstone, Thoma Bravo and CVC before founding Meridian, an AI powered CRM tool for investors. He argued that associates still need to know how to build a model from scratch, and reached for Picasso to explain why.

Before the abstract work, Picasso was an exact draftsman who could paint a technically accurate portrait and chose not to, and that foundation gave the later choices their authority. Partners who have built enough models themselves can see where one is wrong and what a bad assumption looks like buried three levels down. The cost of skipping that stage arrives years later, when the person reviewing the analysis can no longer tell that it's wrong.

AI is now absorbing these formerly learned tasks, and quickly. Datasite and FT Longitude surveyed 1,000 senior dealmakers across 27 countries in March 2026. Sixty-two percent said human-only decision making is no longer defensible in complex transactions and half reported regular AI use in due diligence, but only 22% would let AI make the final recommendation on whether to sign.

These figures describe dissonance across an industry that craves human judgment at the decisive moment while automating the process that has always produced it.

Looking at this through the lens of strategic foresight – a practice focused on making better long-term decisions under uncertainty – the outcome has a ripple effect with the potential to cripple the next generation of investors.

The assumptions built into today's tools will become tomorrow's defaults. What a system treats as a quality signal, a relevant comparable, a credible management team or a reason to pass carries a view about how investing works, encoded by whoever configured the model: a vendor, a data provider or a partner who has since moved to another firm.

A generation trained alongside these systems will inherit these views as truth, without the experience of having formed a competing one. Encoded defaults are also harder to challenge than a partner's stated preferences, which junior investors learn to interrogate and occasionally push back on. You cannot argue with a ranking whose reasoning you cannot see.

Dealmakers are already sensing the stakes. Seventy-three percent report using AI in board reporting and governance, according to the same Datasite research, which puts these systems inside the record of why decisions were made as well as the decisions themselves.

Decision integrity, not decision speed, needs to be the more durable goal for the future of private markets. If firms are going to continue the current pace of AI adoption (which they likely will) it is imperative that the industry start thinking about what AI tools and best practices will optimise the outcome for the next generation of decision makers.

Three recommendations follow from that.

Capture the reasoning behind conclusions. When a firm passes on a company, the record should include what it assumed, what evidence it weighed and what would have changed the answer. Firms revisit the same companies years apart. Most cannot reconstruct why they said no the first time, which leaves them unable to check whether the original logic still holds. Context is critical in an AI-centric world.

Make the model's work inspectable. A system that surfaces what it discarded, and why, teaches the person using it. Most systems return a ranked list and leave the reasoning inside the machine. If the next generation of investors will be AI native, the models that train them should be both the mentor and the assistant.

Design the junior seat deliberately. Change behaviour at its source. Some firms now require analysts to form a view before the model runs, then reconcile the two. While this approach may cost time, doing so ensures a generation of investors that can still think critically, leverage AI tools and produce results.

Picasso earned the right to abandon the rules by learning them first. Today's investors should take note. If not, the private markets industry is at risk of finding out what happens when a generation of investors skip the necessary step of building good judgement through repetition. The firms that get this right will be the ones that stay deliberate about what their people still need to know how to do themselves in an increasingly AI empowered world.