Jo Goodwin
This article expands on a LinkedIn post by Phil Westcott, CEO of Deal Engine.
The infrastructure gap most PE firms aren't talking about
There's a version of this conversation that happens in almost every mid-market PE firm at some point. A senior partner asks why a particular target wasn't on the radar six months ago. Someone checks the CRM. There's a note. There's even a signal, buried in an email thread, or a contact who flagged something at a conference. The information existed. It just never surfaced.
That isn't a people problem. It's an architecture problem.
Private equity has always been a relationship business. Firms built proprietary networks, cultivated intermediaries, and ran thesis-driven origination on the back of individual conviction and hard-won trust. For a long time, that was enough. The advantage was the relationship.
That advantage is narrowing. Every firm now has access to the same data platforms, the same CRM tools, and increasingly, the same AI applications. What separates the firms generating proprietary deal flow from those reacting to what the market brings them isn't access to data. It's what they do with it.
The problem with reactive origination
Most firms monitor their markets reactively. Someone picks up a sector, runs a screen, builds a target list, and checks in on it periodically. Signals that fall outside that window, a revenue inflection, a management hire, a competitor sale, get missed. Not because the firm wasn't paying attention, but because no one can pay attention continuously.
The result is a pipeline that reflects the capacity of your team, not the opportunity in your market. Deals get sourced when bandwidth allows. Monitoring runs on analyst time, not on logic.
This is the constraint that a properly engineered data infrastructure removes.
What the architecture actually looks like
The firms moving from AI experimentation to AI infrastructure are building something specific. Not a new dashboard. Not another data subscription. A unified intelligence layer that sits across every data source the firm uses, external and internal, and runs continuously in the background.
In practical terms, that means three things working together.
- First, data from every source flows into one place. PitchBook, Grata, news and web signals, third-party banker lists, DealCloud or Salesforce, past deal decisions, IC notes, proprietary network contacts, existing watchlists. Not duplicated across systems. Unified, deduplicated, and structured so it can be queried.
- Second, each dealmaker's thesis is codified inside that layer. Not written down in a memo somewhere. Encoded as logic that runs continuously, matching new signals against what each partner is actually looking for. Healthcare services with 20 to 100 million revenue, founder-led, south-east US. That isn't a one-time search. It should be a live filter running against every signal that enters the system.
- Third, the outputs are specific and actionable. Not a list of 400 companies to review. A signal: this target moved to ready-to-transact. This portfolio company competitor just announced a sale. This CFO at a tracked company has changed. These three net-new companies match your thesis parameters as of this morning.
The difference between a tool and infrastructure
Most of the AI spend in PE firms right now sits at the tool layer. A chat interface here, an analysis assistant there. Those tools are useful. But they don't solve the underlying problem, because the underlying problem isn't analysis. It's data fragmentation.
An analyst with access to a brilliant AI model but no unified data stack is still manually pulling from five platforms before they can ask a good question. The AI is only as reliable as the context it has access to. If that context is scattered, partial, or not up to date, the output reflects that.
The firms that will generate compounding advantage from AI aren't the ones with the most subscriptions. They're the ones that have built the layer underneath: the governed, unified, continuously running data engine that makes every AI interaction reliable.
From piloting to infrastructure
According to the AI Pathfinder Private Equity Benchmark Survey from September 2025, nearly half of PE firms are currently in the piloting stage of AI adoption. Only around 11% have reached the scaling stage, and just 2% describe AI as embedded in their long-term investment strategy.
The gap between piloting and scaling isn't a model problem. The models are capable. The gap is data readiness. Without a governed, integrated data architecture, firms cycle through tools without compounding the value of any of them.
Scaling AI is not a tooling challenge. It is an engineering challenge.
What this means in practice
One North American PE firm running Deal Engine across a portfolio of 120-plus employees pushed over 250 net-new deals into their CRM in the first deployment period, with a 99.5% reduction in manual research time. The team didn't get bigger. The infrastructure changed.
The question for most firms isn't whether AI will reshape origination. It already is. The question is whether the underlying architecture is in place to benefit from it, or whether the firm is still running proprietary thesis logic on analyst hours.
Your deal team is brilliant at what they do. The question is what's running in the background when they're not looking.
Want to see what this looks like in practice? Book a demo