How a Mid-Market PE Firm Tripled IC Prep Depth
the challenge
A mid-market PE firm with 20+ years of deal history and a strong IC process wanted to push prep depth beyond what any individual reviewer could realistically deliver. The team did rigorous work, but the depth of any given deal's prep was bounded by one person's bandwidth, one person's recall of past deals, and one person's view of which angles mattered most.
The firm wanted every deal to walk into IC with the kind of prep its best deals had gotten: benchmarked systematically against the firm's 1,000+ opportunity history, stress-tested with the questions senior IC members actually raise, and analyzed through the firm's collective judgment rather than whichever lens a given week's reviewer brought.
what they built
Soal Labs built an IC Insights Engine for the firm: structured interviews with six senior MDs were encoded into queryable AI personas, paired with a knowledge base of the firm's IC decks, deal metadata, and outcomes, so every deal could be cross-referenced, challenged, and benchmarked against 1,000+ past opportunities before committee.
Encoding senior MD judgment into queryable personas and pairing them with the firm's structured deal history meant every deal could be evaluated through the full lens of the partnership and benchmarked against 1,000+ past opportunities.
Soal Labs began with a single managing director. The first two weeks were structured, recorded interviews that walked through past deals to surface how that MD actually reasoned — what they looked at first, which red flags overrode positive signals, and which patterns they had learned to trust. Those sessions became a v0 persona, made retrievable alongside the firm's IC decks and deal data, and were tested immediately on a live deal heading into committee. The next two months were iteration. The naive persona surfaced useful angles but also exposed gaps: judgments too contextual to encode from a single example, and heuristics MDs stated in interview that diverged from what they did when shown a real deal. Each testing round fed back into both the interview methodology and how the personas were structured. Once the approach worked on one MD, the team scaled to five more, deliberately chosen for sector coverage so every vertical had a senior voice encoded from day one. First useful results came within the opening weeks; the full six-persona phase one landed over roughly two months.
best fit for
Most relevant for firms with:
- Significant deal history (hundreds of opportunities with outcomes) for cross-deal benchmarking to draw on.
- Multiple sectors or strategies where no single MD covers everything.
- Senior MDs whose judgment is distinctive and currently lives only in their heads.
impact
5x
6
Osman Ghandour
more from this expert
frequently asked questions
How did a mid-size private equity firm triple the depth of its investment committee prep?
The experts built a decision-support engine that encodes senior managing directors' judgment into queryable AI personas, captured through structured recorded interviews about how each one actually evaluates deals. Alongside the personas, they built an IC knowledge base from the firm's own decks, deal metadata, and outcome records. For each new deal, the system cross-references comparable past opportunities, generates devil's-advocate challenges grounded in the personas, and benchmarks against the firm's history — surfacing far more analytical angles per deal than any single reviewer could.
What AI tools and models powered the IC insights engine?
The work centered on decision support and scoring, built on Claude. Recorded interviews with senior MDs were encoded into queryable AI personas reflecting how each investor thinks, paired with a dedicated IC knowledge base drawn from the firm's decks, deal data, survey responses, and outcome records.
What results did the private equity firm achieve?
Six senior MD personas were encoded in the first phase, preserving decades of judgment, and IC prep surfaced roughly 5x more analytical angles per deal — deeper and broader than any individual reviewer could produce alone.
How long did the IC engine engagement take?
First useful results came within the opening weeks; the full six-persona phase one landed over roughly two months, with iteration on both the interview method and persona structure in between.
Who is this decision-support approach best for?
Firms with significant deal history for cross-deal benchmarking, multiple sectors or strategies where no single MD covers everything, and senior MDs whose distinctive judgment currently lives only in their heads.
Have a similar challenge?
