AI adoption is accelerating across private markets, but moving from experimentation to reliable, institutional use requires more than adding a new tool to an existing workflow.
At SuperReturn CFO/COO, Jessie Juan, Co-Founder of Quantium, and Yann Magnan, CEO of 73 Strings, approached the challenge from different parts of the private markets operating model. Both pointed to the same fundamental issue: before firms can scale AI, they need data they can trust.
Key takeaways
- AI adoption across private markets is growing, but firms remain at very different stages of implementation.
- The complexity of fund structures, IRR calculations and valuations makes accuracy particularly important.
- AI can reduce manual work, but reliable automation depends on clean, structured data.
- As valuations become faster and more frequent, traceability and auditability become increasingly important.
- AI can support judgment, but accountability still sits with people.
Private markets are moving beyond AI experimentation
AI has been a major talking point across private markets for several years, but Juan sees a noticeable change in how firms are approaching it.
At SuperReturn a year earlier, she recalls considerable scepticism. Today, adoption is much more visible, although it remains fragmented.
Some smaller firms are using AI across a wide range of functions. Larger organisations, particularly those sitting on substantial datasets, face a different question: is the underlying information clean and structured enough for AI to use reliably?
That distinction matters because private markets have little tolerance for inaccurate outputs.
A small error can have a big impact
Juan brings an LP perspective to the issue.
A minor numerical difference may appear insignificant in many industries. In private markets, a small change can materially affect an IRR or the information being reported to investors.
Fund structures, allocation policies and the mechanics behind performance calculations add further complexity. That makes simply adding AI to existing spreadsheets a risky proposition.
You can’t simply layer AI on top of Excel spreadsheets and hope that it will produce accurate numbers.
Jessie Juan, Co-Founder of Quantium
The opportunity to automate exists, but Juan argues that firms need to get the foundation right before using AI widely across sensitive reporting processes.
The difficult part isn’t AI
Magnan sees the same challenge emerging in valuations.
Wider access to alternative investments and the growth of structures such as evergreen funds are increasing the frequency at which valuations need to be produced, and changing the consequences of getting them wrong.
A process that once relied heavily on quarterly spreadsheets increasingly needs a more robust infrastructure behind it.
And according to Magnan, the hardest part comes before the AI model itself.
The difficult part is not the AI part.
Yann Magnan, CEO, 73 Strings
The work lies in collecting, normalising, structuring, standardising and consistently labelling the underlying information.
Or, as Magnan puts it:
If you get garbage in, garbage out.
Yann Magnan, CEO, 73 Strings
The foundation comes first
Juan and Magnan approach AI from different parts of private markets, but their conclusions are closely aligned.
AI can automate calculations, accelerate reporting and remove manual work. But none of those benefits eliminates the need for accurate inputs, connected systems, traceable outputs and human oversight.
For private markets, the competitive advantage may therefore come before the AI itself.
Get the data foundation right, and firms have something powerful to build on. Get it wrong, and a more sophisticated model simply produces the wrong answer faster.
