Two funds look at the same hundred companies this year. Similar capital, similar mandates, the same intermediaries returning the same calls. Five years on, one has returned meaningfully more than the other. The deals on offer were identical. The selection was not.
That gap is where private equity alpha actually lives. In a market this efficient, proprietary deal flow – the target nobody else has seen – has largely been competed away. What separates returns now is judgment: which of the opportunities everyone can see are the ones worth backing. And that judgment has a hidden ceiling.
The Ceiling Is Evaluation Bandwidth
Ask a partner what limits the fund and the answer is usually capital, competition, or flow. More often it is none of those. It is how many deals the team can actually get to the bottom of before deciding.
Reaching real conviction on a target is slow work. It means assembling the comparable transactions the firm has done before, establishing how those deals performed after close rather than how they were pitched, and setting all of it against what the market is doing right now. That evidence is scattered across data rooms, IC memos, due diligence files, post-close reports, and external sources, and pulling it into a defensible view takes an analyst the better part of two weeks per target. No pipeline dashboard or M&A pipeline management process changes that, because the bottleneck is the research behind each entry, not the tracking of it.
So the fund evaluates deeply only what it has time to evaluate deeply. Everything else gets a lighter look, or a pass. The ceiling on how many deals you can truly judge is the ceiling on how good your selection can be.
Raise the Ceiling and You Change What You Can Back
This is the real prize, and it is not primarily about saving hours. When conviction takes days instead of weeks, the fund can bring full rigor to several times as many opportunities, which means the better deals are far more likely to get the attention they deserve. In other words, you get more shots, and each one is better aimed. At that scale, efficiency becomes a selection advantage, and selection is the part of the return that compounds.
Speed in deal research is not about doing the same work faster. It is about being able to back a better company than you otherwise would have.
A modern deal sourcing platform raises that ceiling, giving every prospective target a full, 360-degree read against everything the firm has ever learned, in the time it takes to schedule the kickoff call.
Why the Comparable Is a Relationship Problem
The catch is that the deal that rhymes with the one you're staring at is rarely the one that looks like it on paper. A conventional database matches on someone else's lables – sector, size, vintage. It misses the precedent that shares the target's real shape: the same margin profile, the same integration risk, the same post-close trajectory, filed under a different sector entirely.
That is where GraphRAG earns its place. The reason a plain database misses the right comparable is that it stores each deal as a row of fields and can only match deals whose fields line up. A knowledge graph works differently: it stores each deal as a set of connections to the structure it used, the margins it carried, the risks it ran, the way it performed after close, and it links deals to one another through the features they share. So two transactions filed under different sectors and described in different language can still sit a step or two apart in the graph, because they connect back to the same underlying shape.
GraphRAG puts that structure to work. It takes a plain-language question: "What have we done that looks like this carve-out?" and lets a large language model answer it by following those connections instead of matching words. What comes back is the comparable an experienced investor would have reached for from memory: surfaced in seconds, ranked by how genuinely alike the deals are, and carrying its post-close record with it.
Conviction Stays Human
With the comparables in hand, the same system drafts the investment memo, every claim cited to its source. What took two weeks takes hours. It doesn't make the decision: the deal team challenges the comparables, and the investment committee commits the capital. The retrieval and the first draft are automated; the conviction isn't. And every deal researched this way sharpens the next, as the comparable set deepens and the firm's judgment stops living in one partner's memory and starts belonging to the firm.
That same foundation extends past origination into how the portfolio is monitored and how deal terms translate into returns. The discipline underneath – grounded sources, a citation behind every claim – is the same one reshaping regulatory due diligence and the secure information retrieval behind bids across portfolio companies. Only the question changes.
The one question worth asking first: how many good companies did the fund pass on last year, not because they were wrong, but because there was no time to find out they were right?
See how the deal research and origination agent gives every target a full read against the deals you've already done. Every agent in the Squirro catalog is scoped and built around a single business outcome – so you can start where the friction is sharpest and prove the value in weeks. Browse the full Agent Catalog online, or download it as a PDF to share with your team.
