Happy Thursday and welcome to another edition of Beyond TVPI.

For most of my LP career, we had more data than we could afford to use.

AI changes that constraint. Work that once required hours of manual effort is becoming cheap enough to do routinely.

At FundFrame, we are already seeing the benefits of this. This week, I look at one example from portfolio monitoring, and why the opportunity is much larger than making an old workflow 10% or 30% faster.

And, we think the opportunity extends way beyond portfolio monitoring, but that’s a topic for another day.

Enjoy the read,

Steffen

All LPs track capital account values carefully.

Commitment. Paid-in capital. Distributions. NAV. The numbers are reconciled and goes into a report.

To go one level deeper, someone had to open the quarterly report, extract financials for every portfolio company, and repeat the exercise three months later. Company names changed. Currencies differed. Valuation dates did not always line up. Boring, tedious work.

Multiply that by every fund in the portfolio and you had a fairly miserable job. As the screenshot shows, it is 18.5x more data points needed for every fund (for the minimum level of metrics).

That is one of the more interesting things AI changes. Work that was possible but uneconomic (for most) can become routine.

I think we are still underestimating this shift.

Much of the early AI work in private markets - and everywhere else - takes an existing task and makes it faster. Draft the investment memo. Summarise the quarterly report. Extract a table from a PDF.

Useful work. I would happily take the time saving.

But extracting four data points at the capital account level doesn't save much time (and you'd probably want to double-check them anyway). That's the 30% trap.

But if you instead did 18.5x as much work in the same time getting 2-3x the benefit, it is another conversation.

The allocator remains at the end of this process because judgement is where an LP creates value.

Software may show that six companies were written up without corresponding EBITDA growth. The allocator still has to judge the GP's explanation.

The same company may appear through three different funds. The allocator has to decide whether the portfolio is still as diversified as it looked at fund level.

Those are the high-value decisions. They require investment philosophy, experience, calibration and, eventually, accountability when the answer turns out to be wrong.

Those are the things we are solving for at FundFrame.

I wrote recently that judgement is the product. AI makes the point more obvious. Once information becomes cheaper to collect and compare, more of the allocator's time can move towards interpreting it and making the call.

I do not know exactly where the line will settle. Some data will remain too messy. Some analysis will produce detail without insight. Important figures will still need a human review.

But using analyst cost as the reason not to ask a useful question is becoming a much weaker excuse.

I spent ten years wishing we had better information. Looking back, we often had it.

It was just hidden and costly to organize.

—

If you have already tried something like this, or see another workflow where it could create value, I'd love to hear about it. Just reply to this email.

I have been properly knocked out by some kind of illness this week. Probably whatever the latest Covid strain is called.

Three days more or less out of the equation as a founder… stinks. There is always something that needs deciding, answering or pushing forward. When you are out, it waits, or someone else has to pick it up.

That's the way it is in any job. But it does feel like a double whammy when you're ill and also feeling a direct economic impact of it.

This newsletter is written by Steffen Risager, the founder of FundFrame, a platform for LPs to manage their private markets investments.

Before that, Steffen was CIO at Advantage Investment Partners, a Danish Fund-of-Funds.

Steffen has a decade of experience as an LP, and has made commitments totalling approx. $6bn across fund- and co-investments.