The takeaway: The problem is not the AI. The problem is that most firms deploy AI on top of fragmented data, disconnected systems, and teams nobody trained to use it. Until those foundations are fixed, the technology will keep disappointing.

Here is a thought experiment. You buy a sophisticated new analytics platform. You deploy it on top of data that lives in fifteen different spreadsheets, three legacy systems that do not talk to each other, and one person's email inbox. You do not train your team to use it. You do not redesign any workflows. You just... turn it on.

What happens? Nothing useful. The tool is fine. The foundation is not.

This is what most AI adoption looks like in private markets right now. And the results are exactly what you would expect.

The Numbers Are Brutal

MIT's 2025 GenAI report found that despite $30-40 billion in enterprise AI spending, the vast majority produced no quantifiable return.¹ A Gartner survey of finance functions using AI found 91% described early-stage impact as "low or moderate."²

The technology is being adopted. The results are not following.

So what is going wrong? The same thing, over and over.

The Problem Is Almost Always Data

The most common answer from every survey, interview, and industry report is data. Private markets have a data problem that predates AI and makes AI adoption uniquely difficult.

Information is scattered across spreadsheets, PDFs, email threads, and disconnected systems. One study found that 92% of funds of funds experience data quality issues with GP reporting, and teams spend an average of 33% of their time just accessing, processing, and standardizing data — before any analysis happens.³

AI does not fix bad data. It amplifies it. If your systems think "125 Main Street" and "125 Main St" are two different properties, your AI will confidently make recommendations based on that error. Garbage in, confident garbage out.

The firms struggling with AI adoption are overwhelmingly the firms that skipped the boring work of cleaning and centralizing their data first. They wanted the magic without the foundation.

The Integration Problem Nobody Wants to Solve

Half of private equity professionals report that integrating AI with existing systems is a major challenge.⁴ This makes sense when you consider how most fund operations actually work: a CRM here, an accounting system there, a portfolio monitoring tool somewhere else, none of them designed to talk to each other.

Legacy architectures were built for humans to manually reconcile information across systems. Bolting AI onto that infrastructure does not create intelligence. It creates a faster way to propagate inconsistencies.

The Change Management Problem Nobody Wants to Fund

Here is the part that really does not show up well in board presentations: even when the technology works, people do not use it.

Firms buy the software, run a pilot, declare victory, and then watch utilization quietly collapse because nobody trained the team or redesigned their workflows. A recent survey found that 72% of UK private equity respondents still see AI as "more hype than impact."⁵

That is not a technology failure. That is an implementation failure. And it is the most common failure mode.

What the 5% Do Differently

The firms seeing real ROI are not using better models. They are doing the unglamorous work that makes AI actually useful.

They fix the data first. They pick focused use cases instead of attempting whole-company transformation. Document review, data extraction, initial screening — specific, measurable, achievable.

They invest in change management like it matters, because it does.

They treat the first deployment as the beginning of the work, not the end.

Everyone else is buying software and hoping for magic. The magic is not coming.