Measured, not guessed

How does AI answer "best clinic" questions? What we measured in Dubai

Patients don't ask ChatGPT for a clinic's homepage. They ask "best clinic for lip filler in Dubai" or "is CoolSculpting safe" or "how much does a nose job cost in Dubai." We measured what four AI engines actually say back, across 15 real clinics.

28 Labs · September 2026

We asked ChatGPT, Claude, Gemini and Perplexity 80 real patient discovery questions about 15 Dubai aesthetic clinics. The category-average visibility score was 4.5 out of 100. The market leader scored 13.8. Only 2 clinics - Dubai London Clinic and Lucia Clinic - showed up consistently across question types and engines. Everyone else was mostly absent, including some clinics with strong brand recognition among patients.

What patients actually ask AI

Nobody asks an AI engine "tell me about Clinic X." They ask treatment-first, trust-heavy questions: "best clinic for lip filler in Dubai," "is [treatment] safe," "how much does [treatment] cost in Dubai," "which clinic has the best reviews for [procedure]." The question is rarely about the clinic. It's about the treatment, and the clinic is the answer the engine has to supply.

That distinction matters more in medical aesthetics than in most categories. Engines are cautious with health-adjacent recommendations. They lean harder on third-party evidence - reviews, press, listicles, professional directories - because the cost of recommending badly is higher than it is for, say, a restaurant.

What we measured

The Dubai Aesthetic Clinics AI Visibility Index put 80 patient discovery questions to ChatGPT, Claude, Gemini and Perplexity, then checked whether each of 15 Dubai clinics actually appeared in the answers, and where.

80
patient questions tested
15
Dubai clinics measured
4
AI engines: ChatGPT, Claude, Gemini, Perplexity
13.8
market leader's score, out of 100
4.5
category average, out of 100
2
clinics visible consistently, of 15

Visibility varied by engine, too. In our sample, Gemini and ChatGPT surfaced a clinic in roughly 9% of relevant questions, Perplexity around 7%, and Claude the most cautiously at around 4%. No engine treated the category as an easy one to answer.

Who gets named, and who doesn't

Dubai London Clinic and Lucia Clinic were the two clinics AI engines named consistently, across categories and across engines. Kaya Skin Clinic ranked third by mention volume - visible in some question types, not others. The remaining twelve clinics in the index were largely absent, regardless of how well known they are to patients directly.

That gap between patient-facing brand recognition and AI visibility is the finding worth sitting with. A clinic can be a household name in Dubai and still not exist in the answer a patient gets when they ask an AI engine which clinic to choose. The engine isn't rewarding reputation. It's rewarding evidence it can retrieve and trust.

Visible clinics tend to have

  • Treatment-specific pages that name the procedure and the price
  • Deep, consistent review coverage across platforms
  • Coverage in third-party "best of" and comparison lists engines already read
  • Consistent name, address, specialty and pricing facts wherever they appear

Invisible clinics tend to have

  • A homepage and service list, no page that answers a specific treatment question
  • Reviews concentrated on one platform, or thin and inconsistent
  • No presence in the listicles and directories AI engines cite
  • Facts that shift between the website, directories, and social profiles

Why big-name clinics still go unanswered

AI engines answer at the question level, not the brand level. "Best clinic for lip filler" and "best clinic in Dubai" are different questions to an engine, even if a human treats them as roughly the same. A clinic can win the second and lose the first if it never published a page that speaks directly to lip filler - pricing, process, safety, results.

This is why some of the most recognized names in the category scored near the bottom of the index. Recognition is a patient-side signal. AI answers run on retrievable, treatment-specific, third-party-corroborated evidence. Those aren't the same thing, and the index measured the second, not the first.

Where the engines are actually looking

Across the questions we ran, the sources doing the heaviest lifting were the ones any patient would trust too: review platforms, clinic comparison and "best of" listicles, local press coverage, and professional directories. Instagram - where most Dubai clinics post the most - showed up unevenly. It's harder for engines to retrieve and cite reliably than an indexable web page, so an active Instagram account doesn't translate into AI visibility the way it does into patient-side reach. We go deeper on why in does AI read Instagram.

If you want the underlying definitions we use for visibility, mentions and score - Top Recommendation Rate, Visibility Rate, and the rest - they're in the AI visibility metrics glossary.

The honest readA 4.5-out-of-100 category average means most Dubai aesthetic clinics are functionally absent from AI answers about their own treatments. That's not a ranking problem you fix with more marketing spend. It's an evidence problem - the engine has nothing trustworthy to cite you with.

What we do about it

We measure this the same way for every clinic: a fixed set of real patient questions, asked across ChatGPT, Claude, Gemini and Perplexity on a schedule, with every score re-derivable to a specific question and a specific answer. From there the fix is publishable - treatment-specific pages with real pricing, consistent facts across every place a clinic appears, and closing the gap with the third-party sources engines already trust. None of it is guesswork, because the starting point is a measurement, not an opinion.

28 Labs measures how AI engines answer real buyer questions - and what moves the counts. Every number we publish re-derives to specific questions and engines. try28labs.com