The hardest question I have ever been asked in a kickoff meeting came from someone who did not want us in the building.

Twenty minutes into the walkthrough, a director of operations at a large specialty group stopped me. She wanted to know what happens when the agent books a Mohs patient into a twenty-minute follow-up slot because the referral came in mislabeled.

I did not have a clean answer that day. I had a partial one, and I said so.

That practice is now one of the strongest relationships we have. I have stopped treating that as ironic, because it has happened too many times to be a coincidence.

What the skeptic is actually doing

The instinct in a room like that is to read the hard question as an objection to overcome. Somebody isn't sold yet. Bring more proof, more references, more slides.

That reading is almost always wrong, and it costs you the most valuable thirty minutes of the entire implementation.

The director asking about mislabeled referrals was not skeptical about artificial intelligence in the abstract. She was skeptical about a specific failure she has personally cleaned up, more than once, under a different vendor's logo. She knows which referring practices send incomplete faxes. She knows which provider will not accept a certain visit type and has never written that down anywhere. She is asking the only question that matters at that stage: what happens when you're wrong?

That is not an objection. It is a specification.

Every hard question a skeptic asks in week one is a configuration rule you would otherwise discover in week nine, in production, with a real patient in the chair.

The economics of that are not close. A rule surfaced in a kickoff costs a conversation. The same rule surfaced after go-live costs a misbooked appointment, a provider's confidence, an escalation to someone senior, and a week of the ops team wondering what else the system is quietly getting wrong. Trust does not degrade gradually in these deployments. It degrades in specific, memorable incidents.

So the person with the longest list of concerns is usually the person who understands the operation most completely. That is who you want writing the requirements. It just doesn't feel that way in the moment, because it arrives as friction rather than as a document.

The easy yes

The rollouts that have worried me look like the opposite.

Full room. Everyone nodding. Executive sponsor enthusiastic, timeline aggressive, no one asking about edge cases. On paper it is the best kickoff of the quarter.

What it usually means is that nobody in that room actually runs the phones at 4:45 on a Friday. The hard questions have not been answered. They have not even been asked yet. They are still coming, and they will arrive after go-live, from people who were never invited to the meeting — which is the worst possible sequence for both of us, because by then the questions come attached to a live patient and an audience.

Enthusiasm tells you what a client hopes will happen. Skepticism tells you what already does.

What to do when you get one

If you are on the vendor side of this, the skeptic is an asset you have to earn the use of. A few things that have made the difference for us.

— Answer partially and say so. The fastest way to lose an experienced operator is to produce a confident answer to a question you have not thought through. She has heard confident answers before. She is calibrating whether you will tell her the truth in month seven, when something breaks and nobody outside her building knows yet.

— Write the question down in front of her. Literally. A hard question that gets captured as a configuration item, with an owner and a date, converts a skeptic faster than any reference call. It signals that the objection changed something.

— Bring it back with a specific answer. Not "we've addressed that." The rule you built, the safeguard, what the agent now does when the referral type is ambiguous. This is the whole relationship, compressed into one loop, repeated until it is boring.

— Do not route around her. The temptation when someone is difficult in week one is to work through the sponsor who signed. Every time I have seen that done, the deployment reached go-live and stalled at the front desk, because the person who knew where the edges were had been given no reason to care whether it worked.

None of this is a technique for converting critics into fans. It is closer to the opposite. The skepticism does not go away in a good partnership. It gets pointed at the next thing — the new service line, the campaign that isn't converting, the clinic where adoption is lagging. Two years in, the same director is still asking what happens when we're wrong. That question is now the most useful input we get.

The clients who put us through the most in the first month are the ones we have kept the longest. Not because we won them over, but because they told us what the operation actually required before we had a chance to guess.

An easy yes is the most expensive thing a client can hand you.