Expert interviews in due diligence: getting from 10 calls to 100

Expert calls are the backbone of commercial due diligence. When a private-equity firm or a strategy team evaluates a company, the fastest way to understand a market is to talk to the people in it: former executives, customers, channel partners, competitors' alumni. The traditional way to reach them is an expert network — and the traditional constraint is that you'll only ever talk to a handful of them.

How the expert-network model works

The classic workflow: the deal team sends a screening brief to an expert network, the network sources candidates from its roster and cold outreach, the team reviews profiles, compliance clears each expert, and calls get scheduled — usually 30 to 60 minutes, led by a consultant or investor, at a few hundred to over a thousand dollars per call once network fees are counted.

The model works. It has powered thousands of diligences, and a well-run hour with the right former VP of sales can reshape an investment thesis. But its economics and logistics impose a hard ceiling:

• Cost per conversation is high. When each call costs what a working day costs, a diligence budget supports 10, maybe 20 calls. Every call has to justify itself in advance — which means you book the experts you already believe are important.
• Scheduling burns the scarcest resource: deal time. Diligence windows are measured in weeks. Sourcing, compliance, and calendar coordination routinely consume days per expert, and the calls themselves stack up in the final week when the team is busiest.
• Coverage is thin by design. Ten calls across a market with three customer segments, two channels, and a dozen competitors isn't a sample — it's a set of anecdotes. Good teams know this and triangulate carefully, but the ceiling is structural.
• Sequencing prevents learning. The most valuable expert questions are the ones you only know to ask after the first five calls. In a compressed window, calls happen in whatever order calendars allow, so the sharpest questions often arrive after the expert who could answer them.

What short AI voice interviews change

AI-moderated voice interviews don't replace the expert call at the top of the funnel — a 15-minute structured conversation is not a substitute for an hour of expert judgment probed by someone who knows the deal. What they change is everything below that top tier.

Because AI voice agents run interviews in parallel, on the participant's own schedule, at a fraction of the cost, the coverage ceiling moves from "around ten conversations" to "around a hundred." That enables a different diligence structure:

1. Screen wide first

Run short, structured voice interviews across a broad slice of the market — customers of the target, customers of competitors, former employees, channel partners. Fifteen minutes is enough to test the thesis's key claims: Why did you choose this vendor? What would make you switch? How did the renewal conversation actually go? Every interview follows the same guide, so answers are comparable across the whole set.

2. Let the synthesis find the signal

With 80–100 structured conversations, patterns stop being anecdotes. If churn risk concentrates in one segment, or three ex-employees independently describe the same product weakness, that shows up as a pattern with traceable quotes behind it — not as one memorable call that may or may not generalize.

3. Spend human hours on the calls that earned them

The wide screen doesn't eliminate premium expert calls; it tells you which ones to book. Instead of guessing which ten experts matter before you understand the market, you enter the expert-network process knowing the three questions that actually decide the thesis and which profiles can answer them. The expensive hours get spent probing the highest-value uncertainties, with the sharpest questions — the ones that used to arrive too late — prepared in advance.

The honest limits

This structure has boundaries worth stating plainly. Very senior operators — the former CEO, the divisional president — warrant human-led calls, full stop, and some will not do an AI interview at all. Compliance rules around MNPI and conflicts apply to AI-moderated conversations exactly as they do to human ones, and screening still matters. And a 15-minute interview yields breadth, not the improvisational depth a skilled diligence interviewer extracts in an hour.

Diligence has always been about reducing uncertainty per unit of time. The teams that cover the market widely before going deep aren't doing different diligence — they're doing the same diligence with the coverage it always wanted.