Future LP Workflows

How different will allocator life look post AGI?

Below is an illustration of what the day-in-the-life of a Limited Partner could look like post-AGI. Of course, there are reasons every item discussed could never happen, but note that they are cultural and policy reasons, not technical reasons, which means they will be under our control.

Let’s dig in!

---

It’s Friday morning (date unknown). You have an intro manager meeting at 9:00 AM.

To make the best use of their time (and ours), managers have started giving us access to their data rooms upfront. These are “intro-meeting-approved” data rooms — sensitive info is redacted, but the basics are there.

In this post-AGI future, these data rooms are always up and running, not stale or only refreshed during fundraising. That’s because:

• Lower Cost, Higher Value: Agentic automation collapses the marginal cost of keeping materials current. Back-office systems produce increasingly automated valuations, holdings, and portfolio updates—GPs simply need to allow access to the information they have readily available (redacted as needed). Today, humans can’t process and utilize this volume of data; tomorrow, an LP's AI agents can turn this data into valuable insights.

• Resource Efficiency: Live data rooms eliminate bespoke Q&A and repetitive requests, freeing GP teams from manual work.

• Competitive Pressure: As LPs see the benefits of real-time data (e.g., transparency, benchmarking, market mapping, etc.), GPs that maintain live data rooms gain an edge, while those who don’t risk being deprioritized (especially in market mapping).

With access to a data room (or at least detailed pitch decks), you leverage an AI agent to draft an investment recommendation (“IR”) before the intro meeting. You’re not walking in blind… you’re walking in with a near-complete IR for every intro meeting.

If the GP’s data room is missing key info you usually ask for, your AI agent will autonomously reach out to the manager’s AI agent to auto-fill things like GP commit, number of portfolio companies, etc.

Because you’re now swamped with useful materials to review ahead of every intro meeting, (i) you’re only taking meetings with GP’s whose AI-created IR scores high enough on your scorecard, and (ii) your digestion of information has to change.

On your drive to work (RIP remote work 😉), your AI generates a podcast-style summary of all the meetings you have that day: your 9:00 AM intro meeting with the GP, your 1x1 with your boss, and your legal terms discussion – the priming for all of these meetings is provided by Morgan Freeman’s soothing, AI-generated voice.

But it’s not a monologue. You can interrupt Morgan and ask him key questions:

• Who else is executing a similar strategy?

• What does that legal term actually mean, and what percent of the time has your firm accepted it in the past?

• What were the areas of improvement you told my boss you’d work on?

Morgan will respond in real-time. You get a fully interactive, dynamic audio briefing on the go.

You now walk into the meeting at 9:00 AM.

Morgan and your AI agent have prepared a detailed question list, customized to areas that weren’t fully clear during your drive-in briefing.

As the meeting progresses, the agent adds and removes questions on the fly, based on what’s covered. At 9:55 AM, the two questions you need to complete the IR are automatically pulled to the top of your question list.

Importantly, they won’t be basic questions like: “What percent will be invested in Asia?” or “What will be the target equity check?” Your questions will focus on the things your agent AI can’t answer:

• Is this person likable?

• Do I believe they’re differentiated?

• Are they really committed?

• How do they think through hard problems?

As you leave the meeting, you click one button and go grab a coffee.

If you click “Decline”, your agent drafts a polite decline email that is scheduled to send 5 days later.

If you click “Keep warm”, your agent drafts a thank you message and updates your pipeline.

If you click “Interesting”, your agent updates the memo to reflect the most recent call and produces a final draft of the IR. Frankly, this IR is way better than what you use today:

• More detailed

• More accurate

• Easier to read

• Every key point links back to the original source (data room or call transcript)

It’s now 10:15 AM, and you’re reviewing the IR. Importantly, you’re reviewing the document while the meeting is still fresh in your mind. No more struggling to remember things from half-baked notes created 3 months and 146 meetings ago.

As you review, if you’re not sure where a comment or datapoint comes from, just click it — it’ll take you to the exact moment in the conversation or the supporting file.

Since you clicked “Interesting” as you walked out of the meeting, an AI agent automatically checked your trusted LP network (10–20 peers who have mutually approved each other) to see:

• Have they met with this manager recently?

• Have they opted into a reference call on this name?

If yes, a reference call is auto-scheduled (a similar but upgraded version of what Calendly can do today).

If someone’s free, you could be on the call by 11:30 AM. And that call… It’s all about what AI still can’t assess:

• How likable is the GP?

• How do they handle adversity?

• Do they have real grit?

The AI listens in, takes notes, and folds the insights back into the IR.

It’s now lunchtime… and as you eat your Sweetgreen, you ponder the recommendation. You ask a few clarifying questions to your Morgan Freeman-voiced AI agent and then mutually agree it’s worth bringing to investment committee.

It’s 1:00 PM - 4 hours after you met this manager for the first time and all you have to say is:

“Post the IR for investment committee discussion on Monday.”

---

The story above transforms a 4-24-month process into a 4-hour process. Again, there are reasons every item discussed above could never happen. But, again, they are cultural and policy reasons, not technical reasons.

Historically, intelligence has been a key driver of alpha. Now, intelligence is being commoditized. Going forward, it is the LPs who navigate the cultural and policy challenges of utilizing intelligence that will win. And importantly, they will win via speed and quality.

The point of the story above is not simply to demonstrate that an LP can go from meeting a manager to creating an IR in 4 hours. This story illustrates that an IR completed in 4 hours post-AGI could be a cleaner, better-informed, and overall higher quality work product than the ones we’re currently making today on the back of a 4–24-month process.

To repeat, 4 hours of AGI-aided work will produce a better work product than the 4-24-month processes we’re currently running.

This is the level of disruption coming to our industry.

This is the level of disruption coming to every industry.

In my view at least… I’d love to hear where you agree and disagree!