73% of New York plastic surgeons are invisible to AI.
We asked ChatGPT, Gemini, Perplexity and Claude the questions patients actually ask — "best plastic surgeon in New York," "who should I hire" — for 103 New York-area plastic surgery practices. Then we recorded whether each engine named them. The result: most practices don't exist as far as AI is concerned.
Published July 2026 · being re-validated — please don't cite the figures until the refreshed edition is out.
Four engines. 103 practices. One verdict.
Every engine ignores the vast majority.
This isn't one bad engine. All four leave out roughly nine in ten practices — the exclusion is systemic, not a fluke of any single model.
A few names take almost half the recommendations.
When AI did recommend someone, it recommended the same handful over and over. Just three names accounted for 46% of every recommendation across the whole sample — out of 40 distinct names that ever appeared. AI answers reward incumbents; breaking in is the whole game.
The engines don't agree — which is the good news.
On 27% of practices the four engines disagreed: one would name a practice the others ignored entirely. If AI visibility were fixed, they'd all say the same thing. They don't — which means the signals engines read can be changed, and a practice invisible today can be surfaced tomorrow.
How we ran it — and what it does and doesn't claim.
- Sample: 103 New York-area plastic surgery practices scanned in July 2026. This is a real convenience sample from our prospecting list — not a random survey of every NYC practice.
- Method: each practice was run through The AEO Loop scanner — three real buying-intent prompts ("best plastic surgery in New York," "who should I hire," "top recommended providers near New York") against ChatGPT, Gemini, Perplexity and Claude.
- Scoring: a neutral cross-model judge classified every answer into one of four states — Recommended, Mentioned, Cited, or Excluded — and mapped it to a 0–100 visibility score. No engine graded its own output.
- Caveat: AI answers are non-deterministic and change over time. Read these figures directionally, as a snapshot — not a permanent ranking.
Being great isn't the same as being recommended.
Most of these are excellent, established practices with strong websites and real reputations. They're still invisible — because AI reads different signals than patients do. The fix isn't more ads; it's making your practice legible and trusted to the engines. That's what AEO does.
About this study.
How many NYC plastic surgery practices are invisible to AI?
In our July 2026 study of 103 New York-area plastic surgery practices, 73% were excluded from all four major AI engines — ChatGPT, Gemini, Perplexity and Claude — when we asked the buying-intent questions patients actually use. Only 13% were recommended by even one engine.
Which AI engine recommends the most plastic surgeons?
None recommended many. Gemini was the least exclusionary but still left out 86% of the practices we tested; ChatGPT excluded 91%, Perplexity 90%, and Claude 89%. The average practice scored 16.7 out of 100 for AI visibility.
Why are so many plastic surgeons invisible in AI answers?
Because AI engines recommend the businesses whose signals they can read and trust — structured, extractable content plus citations in the sources engines draw on. A strong website and reputation are not enough on their own; without those machine-readable signals, the engine confidently names a competitor instead.
Do the AI engines agree on who to recommend?
Often not. The four engines disagreed with each other on 27% of the practices we tested — one might name a practice while the others ignored it entirely. That inconsistency means AI visibility is not fixed; it can be influenced with the right work.
How was this study conducted?
We used The AEO Loop's scanner to ask ChatGPT, Gemini, Perplexity and Claude three real buying-intent questions about each of 103 New York-area plastic surgery practices in July 2026, then classified each answer into one of four states — recommended, mentioned, cited, or excluded — with a neutral cross-model judge. It is a snapshot; AI answers shift over time and should be read directionally.
Is your practice in the 73%?
Run the same scan we used for this study, on your own practice, free. You'll see your verdict across all four engines in about a minute — and a founder-reviewed Gap Report within 24–48 hours.