Get your free snapshot

GEO / how to

How to check whether AI engines are recommending your competitors

Open ChatGPT, Perplexity and Google's AI mode in a logged-out or private window, ask the five questions your patients actually ask before booking, and write down every practice name each engine returns. Run each question twice, in separate sessions. The names that come back repeatedly are the practices currently winning your market inside AI answers — and if yours is not among them, you now know precisely which questions you are losing.

Hands at a laptop with an AI analytics dashboard of charts and metrics floating above the keyboard

This takes about twenty minutes and requires no tools. It is worth doing yourself at least once, because the result tends to be more persuasive than any report someone hands you.

Step 1 — Write down the five questions patients actually ask

Not the keywords you track. The questions. The phrasing a nervous patient types when they are deciding whether to spend several thousand dollars on their teeth.

For most cosmetic and specialty practices they look like: who does the best veneers in [city]; how much do dental implants cost in [city]; what is the best cosmetic dentist near [neighbourhood]; is [procedure] worth it; who should I see for [specific problem] in [city].

Use your actual city and your actual procedures. Generic versions will give you generic answers and teach you nothing about your market.

Step 2 — Use a clean session

This is the step people skip, and skipping it invalidates the whole exercise. If you are logged in, the engine may know your location, your history, and — if you have visited your own site repeatedly — something about your practice. You will see a flattering answer that no patient would ever get.

Log out. Use a private window. If the engine asks for a location, give it your city rather than allowing precise location access. You are trying to reproduce a stranger's result, not your own.

Step 3 — Ask each engine and record the names

Run all five questions through each engine in turn. For every answer, write down two things: which practices are named, and what reason the engine gives for naming them. That second column is the valuable one — it tells you what the engine believes each practice is known for, which is a direct read on what its sources say.

Note the cited sources too, where the engine shows them. You will usually find the same few directories, association pages and review sites doing most of the work. Those are the surfaces that decide your market.

Step 4 — Read the pattern, not the single result

Generative answers vary between runs. One appearance is noise; appearing in four of six runs is a position. This is why you ask twice in separate sessions, and why a single screenshot of a good answer is not evidence of anything.

Tally how often each practice appears across all your runs. What emerges is a rough leaderboard for your market — and usually one or two names that dominate far more than their size or reputation would suggest.

What the results actually tell you

If a competitor appears constantly, look at what the engine says about them and where it got that from. It is rarely because their website is better than yours. It is almost always because something outside their website — a directory profile, a detailed review presence, a local article, a professional listing — gave the engine something checkable to repeat.

If no practice is named consistently, the market is unclaimed. That is the best possible result to find, and the shortest window to act in.

If you appear but the reason given is wrong — the engine describes you as a general dentist when you are an implant specialist — that is a structured data and positioning problem, and it is usually the fastest of all of these to fix.

Where this stops being a do-it-yourself job

The manual version tells you where you stand today. What it cannot do is tell you whether you are moving, because the answers shift as engines re-crawl and competitors publish. Tracking that means running the same question set on a schedule and recording it consistently over months.

That is the part worth handing off. But run it yourself first — twenty minutes and a notepad will tell you more about your actual competitive position than most marketing reports you have paid for.

Or have us run it properly

The free visibility snapshot does exactly this across ChatGPT, Perplexity and Google AI, with a full question set and named competitors, and returns it as a report you can act on. It is free, and it does not require a call.

Get your free snapshot
← All articles