Why doesn't ChatGPT recommend my aesthetics clinic?

Clare Gallivan · 7 September 2026 · 6 min read

Open ChatGPT and ask it for the best aesthetics clinic in your town. If a competitor's name comes back and yours doesn't, that isn't bad luck — and it usually isn't about how good your treatments are. It's about whether the machine can read you.

Every clinic owner I speak to has noticed the same shift. Clients arrive already decided. They mention "I asked ChatGPT" the way they used to say "I googled you". A quiet chunk of your market is now being filtered before anyone reaches a search results page — and unlike Google, there's no page two to be found on. An AI assistant names two or three clinics, and everyone else is invisible.

Here's what's actually happening, and what to do about it.

First: how an AI assistant decides who to name

Language models don't rank pages the way Google does. When someone asks for a local clinic recommendation, the assistant is doing some combination of three things:

• Recalling what it absorbed during training — which favours businesses written about in press, guides and "best of" round-ups.
• Retrieving live pages through a crawler, then summarising what it finds.
• Reconciling whether the business it found is definitely the one being asked about — name, location, services, contact details all agreeing.

Fail any one of those and you don't get named. Most clinics fail at least two, and never find out.

The six reasons your clinic isn't named

1. AI crawlers literally can't read your site

This is the most common and most fixable. Site builders and security plugins routinely block or throttle automated visitors — and the crawlers used by ChatGPT, Claude and Perplexity look automated because they are. I audited a permanent-makeup studio recently whose website was returning "410 Gone" to non-browser visitors: a code that tells any crawler this page has been permanently deleted. Their content was excellent. No AI tool could see a word of it.

Check your robots.txt, check whether your sitemap actually loads, and check that your key pages return content to something that isn't a browser.

2. Your content only exists inside JavaScript

Plenty of modern site builders render treatment pages and blogs client-side. Humans see a beautiful page; a simple fetch sees an empty shell. If your treatments only exist after scripts run, assume the machines aren't reading them.

3. Your business facts disagree with each other

Your Google profile says one name. Your website footer says a slightly different one. Fresha has an old phone number, Treatwell has a duplicate listing from three years ago, and your postcode is abbreviated in your site's code. Every mismatch lowers a model's confidence that these are all the same clinic — and a system that isn't confident stays quiet rather than risk being wrong.

Pick one canonical name, address and phone number. Make every platform agree, exactly, down to the punctuation.

4. There's nothing quotable about you

AI answers are built from sentences that can be lifted. "Award-winning treatments in a luxury setting" can't be quoted usefully — it says nothing. "Lip blush at our Clapham studio takes about two hours, costs from £X, and needs a top-up at six to eight weeks" can. Clinics that publish clear, specific, question-shaped content get quoted. Clinics that publish atmosphere don't.

5. Nobody else mentions you

Models lean heavily on third-party sources: local press, "best clinics in…" listicles, directories, community threads. If your clinic appears nowhere except its own website and its booking platform, there's little corroboration to draw on. This is the slowest of the six to fix and the most valuable — one inclusion in a genuine round-up can outlast a year of social posts.

6. You have no llms.txt

An llms.txt file sits at the root of your site and states plainly, for machines: who you are, where you are, what you treat, what it costs, how to book. It takes half an hour to write. Almost no clinic has one, which currently makes it a cheap advantage.

What to do first

In order of impact for a typical independent clinic:

• Unblock the crawlers and repair your sitemap. Nothing else matters until the machines can read you.
• Make your name, address and phone identical everywhere, and merge duplicate listings.
• Add structured data — including your review rating — so search engines and assistants can parse who you are and how well regarded you are.
• Publish llms.txt.
• Write eight to twelve genuinely useful question pages: cost, healing, aftercare, comparisons between treatments.
• Get named somewhere that isn't yours: local press, a round-up, an interview.

The clinics that win here won't be the ones with the biggest budgets. They'll be the ones whose information is clean, readable and consistent — because that's what an AI assistant needs before it will say your name out loud.

Common questions