The Human Premium

Something new has been sitting in on every healthcare sale process we've touched this year: artificial intelligence. Buyers use it to read data rooms overnight. Lenders use it to screen financials. Sellers use it to polish their story. Much of this is genuinely good — and some of it is quietly dangerous, because a growing number of buyers are letting the machine do the part of diligence that was never about documents: understanding the business, and the people who built it. This issue is about where AI truly earns its seat in a sale, where it has almost nothing to say — and why, in a market full of instant analysis, the premium is moving to something older: trust, experience, and the person across the table.

What changed in the buyer's chair

Ten years ago, a buyer's first pass at your company took a team and a month. Today a two-person shop can upload a confidential information memorandum and three years of financials and have a summary, a question list, and a draft model before lunch. Nowhere is this more visible than among independent sponsors — deal-by-deal buyers who run lean, raise capital per transaction, and are among the most active acquirers of founder-owned healthcare businesses today. They also lead the middle market in adopting AI diligence tools: document scanners that flag change-of-control clauses, models that hunt anomalies in the ledger, assistants that generate hundred-item request lists in an afternoon. None of this is bad. But it has produced a new species of buyer: one who has "done the work" without ever having understood the business — whose conviction lives in a model, not in a relationship. And conviction that lives in a model dies in a model, usually at the worst possible moment of your deal.

Where AI genuinely earns its seat

Let's be fair to the machines first, because owners should want them in the deal — on both sides of the table. AI is excellent wherever the answer is already written down:

Reading at scale. Scanning hundreds of payor and customer contracts for assignment and change-of-control clauses, termination rights, and rate resets — in hours instead of weeks, with fewer misses than a tired associate.

Finding the anomaly. Trend breaks, margin drift, seasonality that doesn't repeat, add-backs that don't tie — pattern detection across years of ledgers is real, and it's now table stakes in quality-of-earnings work.

Preparing the room. Completeness checks, indexing, first-draft disclosure schedules, Q&A logistics. Weeks of process friction, compressed.

And for you, the seller: rehearsal. The smartest owners now run the buyer's tools on themselves — let the machine find your gaps, your inconsistent numbers, your missing consents — a year before any buyer does. Findings you fix are footnotes; findings a buyer discovers are discounts.

Where the machine has nothing to say

Here is the problem: almost everything that decides whether a deal closes well is not written down anywhere. Why your referral sources actually send you patients — the twenty-year relationships, the trust built visit by visit, the loyalty a model reads as "at-will." Whether your clinical director will stay after close — and what your best people will do if the culture turns. Whether your forecast is conservative or hopeful — a judgment made by looking a founder in the eye across three meetings, not by regression. What you, the owner, actually need from the deal — the structure, the timeline, the legacy terms that make you a motivated seller at one price and a reluctant one at another. And above all: whether the promises in the agreement will be honored. An earnout is a bet on a person. A seller note is a bet on a person. A transition period is a bet on a person. No model can diligence a promise.

The caution flags — and what they predict

From the sell side, you can now spot the buyer whose conviction is machine-made, and it matters, because the pattern is predictable. Watch for the buyer who has "completed preliminary diligence" without ever meeting management — who makes an offer shaped entirely by your data room. Watch for the question list that wasn't written for you — generic, slightly off-industry, asking a home-health agency about inventory turns or a behavioral health group about foot traffic. And watch for what happens at confirmatory diligence, when reality diverges from the model — as it always, always does. The buyer who invested in understanding you absorbs the surprise, because their conviction has more than one leg. The buyer who outsourced understanding to software re-trades — late, at the moment of maximum leverage, when your exclusivity has run and your alternatives have gone cold. With independent sponsors there's a second-order version: their capital providers re-run the diligence, and if the sponsor's knowledge of your business is an inch deep, that financing wobbles exactly when you need it firm. The headline number on an LOI from someone whose model knows you is not worth the number on an LOI from someone who knows you.

The owner's checklist

Use the tools yourself, first. AI-audit your own data room before anyone else does. Arrive machine-proof.

Never let the machine speak for you. Insist on management meetings before an LOI is signed — and win them. The buyer's conviction about you is the asset their model can't produce and their investment committee most wants.

Ask every buyer one question: "What do you want to understand about this business that isn't in the data room?" A thoughtful answer reveals a real buyer. Silence reveals a model wearing a suit.

Reference them like they reference you. Call owners who sold to this buyer. One question: how did the closing terms compare to the LOI? That delta is the buyer's character, in numbers.

Weight certainty like price. A number is an opinion until it survives confirmatory diligence, financing, and the buyer's own backers. Experienced advisors carry the information no tool has: who closes at their number, and who negotiates twice.

None of this is nostalgia. We use these tools daily and expect every serious buyer to do the same — the machines have made deals faster, cleaner, and better documented. But a sale is not a document exercise. It is a founder handing over a life's work on a set of promises, and promises are kept by people. AI shortened the paperwork. It did not shorten the distance trust has to travel. In this market, that distance is where your premium lives.

This brief is general information drawn from our transaction experience, not investment, legal, or tax advice. Every process differs; take decisions with your own advisors at the table.