The Single Point of Failure
Every private equity firm has one person who "knows everyone." That person is incredibly valuable, and also a single point of failure.
Ask them about a sector and they'll remember the founder they met at a conference four years ago, the operating partner who'd be perfect for a board seat, the deal that looked just like this one and fell apart.
A new deal lands and now someone asks: who do we know who really understands this space? Over six years the firm has probably met fifty experts. The team will remember about twelve, and most of those are people they met in the last year or two. Go back further and the rest are simply gone. The emails still exist. The call notes still exist. The CRM still has the names. But nobody can pull it all together in the moment they actually need it.
That is a plumbing problem. And it's why more funds are building a knowledge graph, and why it's quietly turning into basic infrastructure for private equity, the same way a CRM did twenty years ago.
What Is a Knowledge Graph?
It is a map of how everything in your firm connects.
Normal company tools keep information in separate boxes. Your CRM has data about companies, your inbox knows about conversations, your shared drive knows about documents. None of them know about each other. So when you want the full picture on a company, you (a human) have to open five tools and stitch it together in your head.
A knowledge graph does that stitching for you. You feed in everything: documents, emails, meetings, the CRM. Then it links it all around the two things a fund cares about most, people and companies. Now "this company" is connected to "the founder," who is connected to "the partner who met him," who is connected to "the similar deal we passed on last year." You follow the threads instead of digging for them.
This isn't a startup fantasy. Microsoft's own research team built a version of this idea (they call it GraphRAG) and found that ordinary AI search "struggles to connect the dots" when the answer is spread across many documents. The core idea behind a graph is that it explicitly maps relationships between pieces of information so those connections can be discovered later.
Private equity is exactly that kind of problem. The answer rarely sits inside one memo. It's usually scattered across call notes, banker decks, CRM records, operating reviews, emails, and expert interviews. A graph is built to connect those dots.
Siloed tools
CRM, inbox, and drive each hold one slice of the truth and none of them talk to each other, so the full picture only ever exists inside a person's head.
The AI Brain
60x builds one of these for each firm and calls it the AI Brain. It turns siloed company data into a searchable, always-updating system that maps people to projects, documents to decisions, and communications to outcomes.
The name is the whole point. It's not a search box. It's closer to a colleague who has read everything the firm has ever produced and never forgets any of it.
From a Messy Inbox to a Scored Deal
The easiest way to see it is the banker deck.
A bank emails over a one-pager: a grid of company logos, EBITDA along the top, sector down the side. Here's how the firm used to handle those. A person copied the companies into a spreadsheet. Pasted them into ChatGPT. Copied the answers back into the spreadsheet. Moved the spreadsheet into another tool. Exported it. Then typed each company into the CRM, one at a time. And half the time the deck just sat in someone's inbox and never got processed at all.
The same job with a knowledge graph behind it can improve the entire process. The deck arrives and gets read automatically. Every company in it gets pulled out and looked up: the last ten years of history, who's joining or leaving, whether they're hiring engineers or cutting them, whether they switched from a services business to a software one. Then the system scores each company against what the fund actually likes to buy, and only the ones that pass get added to the CRM.
The old way ate a morning and quietly lost the decks nobody opened. The new way runs the second the email lands, every time, and never drops any information.
Why Start With Finding Deals
A fund could point this at lots of things. The best place to start is deal sourcing.
Portfolio reports and investment memos are easy wins too. But deal sourcing is the right first target, for two simple reasons.
It's where the hours go
Partners at one firm said finding deals was the single biggest use of their time, because a fund that does only a few deals a year spends most of its life at the top of the funnel, looking.
It's the hardest part to copy
Anyone can run a financial model. Knowing which deal is worth chasing is the real skill. You don't want AI to replace that judgment. You want it to feed that judgment by putting more good options in front of the partners.
This lines up with what Bain's Global Private Equity Report has been saying for years: the firms pulling ahead are the ones with an edge in proprietary deal sourcing. AI is now showing up across the deal process, but the biggest advantage may not be finding entirely new companies. It may be remembering every company, founder, advisor, banker, and expert the firm has already encountered.
One firm reported that after putting this in, the number of deals it could screen in a year went from roughly 500 to roughly 5,000. Not to do the same work for less money, but to look at far more targets and pick better ones.
Don't We Already Have ChatGPT or Copilot?
Fair question. The honest answer is no, it's not the same thing.
Tools like ChatGPT and Microsoft Copilot are great at "what's out there in the world?" They're weak at "what do we know?" The answer to that lives in your private files, your old emails, your call notes. None of that sits on the public internet, so a general tool simply can't see it.
That's the whole gap. A knowledge graph reads your private, internal stuff, the part no public tool can reach. So the two actually work together: the graph writes a strong first draft from what the firm knows, and you polish it in Copilot before it goes out. One feeds the other.
There's also a quieter reason the cheap option disappoints. Wiring a few tools together so it works once is easy. Making it still work in six months, every day, as new data keeps pouring in, that's the hard part. And that's what makes this infrastructure instead of a gadget.
Is This Just a Way to Cut Costs?
No, and treating it that way gets private equity backwards.
PE is a high-margin business. Funds aren't desperate to shave a few salaries. So the win here isn't "cheaper," it has to be "better decisions." More deals looked at means better targets reach the table, which means better investments. A single good investment is worth far more than a saved cost, and it keeps paying back for years.
The goal is to make the firm smarter and bigger. Judge the tool by money saved and you'll under-invest in it. Judge it by deal quality.
How Long Until It Works?
Weeks, not quarters. And the hard part isn't the technology.
Something like this should take a few weeks and can be ready in less time than most people expect. 60x says it can stand up a working version on your own data in about 14 days, with the typical path to payback inside roughly six months.
The technology is the easy part. The hard part is people. You have to agree on where to point it first, and the team has to get comfortable trusting a draft the system wrote instead of waiting on the partner who "remembers everyone." The real work is helping people change their workflows and adapt to these tools to get the best results.
Bottom Line
The deal-sourcing edge a firm is proudest of is often sitting inside its own files right now.
The emails, call notes, expert interviews, conference meetings, banker introductions, and investment memos already exist. The problem is that most of that knowledge becomes harder to access every year.
For years, proprietary sourcing was about building relationships. Increasingly, it may also be about remembering and reusing those relationships better than everyone else. The firms that win won't necessarily be the ones with the most information. They'll be the ones that can connect it, search it, and apply it when the next opportunity appears.
A knowledge graph is one way to make that happen. It turns institutional memory from something locked inside a few people's heads into something the entire firm can use in seconds.