The machine doesn’t know her
Premium brands are automating a relationship they never built the record for.
Issue 10 of The Retention Edit
A founder I work with forwarded me a pitch last week. AI clienteling, the subject line promised. The deck was beautiful. Predictive rebooking. A model that “knows when each client is ready to return.” Personal, automated, at scale.
She asked me one question.
Should we buy it?
It’s the right question, asked slightly too late in the sentence. Because the thing being sold here isn’t really software. It’s a promise — that a machine can hold the relationship a clinic keeps losing in the last ninety seconds of an appointment.
It can’t. Not yet, and not the way the deck implies.
Here is what the brochures leave out. AI in clienteling is only ever as good as the data sitting underneath it. Feed it a rich, current, honest record of a client — what she came in for, what she’s anxious about, when she last returned and why — and it can prompt the right person to reach out at the right moment. Feed it the record most premium clinics actually keep — a name, a treatment code, a half-finished note — and it does something quietly expensive.
It automates the gap. It guesses, confidently, at scale.
The luxury operators who have gone furthest with this technology tend to say the same thing in private. The model only works with a person in the loop: the machine suggests, the human decides. The data has to be connected before the intelligence means anything. Switch the AI on over a broken process and you haven’t fixed the process. You’ve made it faster, and harder to see.
This matters more in aesthetics than almost anywhere, because here the relationship is the product. A client doesn’t return for the treatment alone. She returns because she felt known. No algorithm manufactures that feeling. At best, it protects the conditions for it.
So the question is never should we buy the AI. It’s what does the AI have to work with.
With one clinic, we didn’t start with software at all. We started with the close. Those ninety seconds after the treatment, where rebooking either happens or quietly doesn’t. We rebuilt what gets captured in that window: not only the next appointment, but the reason for it in the client’s own words. The concern she’d named. The date she should sensibly return by, for that treatment. The person on the team she trusts.
One change — disciplined capture at the close — did two things.
It lifted rebooking before any technology touched it. And it gave the eventual automation something true to act on. When the reminders did go out, they landed, because they were built on a real record rather than a guess dressed as one.
Strategy, then system, then software. Reversed, it’s just guessing — faster.
This is the work Tenue does before a clinic spends a pound on AI. We fix what the relationship is built on: what you capture, when, who owns it, and how it carries from one appointment to the next. The technology comes last, because by then it finally has something to be intelligent about.
If you’ve been sent one of those beautiful decks — or you’re already running a tool that hasn’t quite delivered what it promised — that’s usually the tell. The software isn’t the problem. The foundation under it is.
That’s where we’d start.
Until next time,
Alesia
Founder Tenue Consulting