Somewhere near you, right now, a person is asking ChatGPT where they should train. The answer comes back as a short list of names. Either you are on it or you are invisible, and most gym owners have never once looked.
This is the five minute version of the test we run professionally. Grab your phone, open ChatGPT or Perplexity, and find out where you actually stand.
The four questions to run
Ask the questions your next member would ask, in their words. These four shapes cover most of the ground:
- “What is the best gym near me?” Run it with your area name if the assistant does not know your location, for example “best gym in Kreuzberg”. This is the broadest contest and the hardest to win.
- “What is the best gym in [your city] for beginners?” The beginner qualifier changes everything, because now the assistant is hunting for signals that you welcome new people.
- “Which gyms in [your neighbourhood] have classes?” Class questions test whether your timetable is readable. Many gyms lose here for a boring reason: the schedule lives in a booking widget that machines cannot open.
- “Can you recommend a personal trainer in [your city]?” If you have coaches, this is their contest. Assistants match named trainers with described specialties to the goal in the question.
These are simplified versions of the patterns we broke down in what people ask AI before joining a gym. If you want the full picture of question types, read that piece as the companion to this one.
How to run them properly
A little discipline makes the result trustworthy.
Use a fresh conversation for each question, so one answer does not colour the next. If you are logged in, remember the assistant may know your history; the cleanest read comes from asking the way a stranger would, with no hints about your own gym.
Run each question twice, once in ChatGPT and once in Perplexity. They work differently under the hood and one may know you while the other does not. That difference is itself useful information.
Then ask the follow up that owners always skip: “Why did you recommend those?” The assistant will usually explain what it found. Listen closely, because it is telling you which sources it trusts about your area.
How to read the answers
There are three outcomes for your gym, and each one tells you something different.
Named and described well. You appear, the description is accurate, the hours and specialties are right. This is the goal. Note the wording the assistant uses, because that wording came from somewhere, and you want to keep those sources current.
Named but vague or wrong. You appear, described as “a local gym” with no detail, or with an old address and dead class times. Close, and still losing. A prospect comparing you against a rival described in confident, specific terms will pick the rival.
Absent. The assistant names your competitors and behaves as if you do not exist. This is the outcome most independents get, and it has identifiable causes.
What absence means
When your gym is missing, it is nearly always one of four patterns, and you can tell them apart.
Blocked AI crawlers. Some websites block AI crawlers in robots.txt without their owner ever knowing, often via a platform default or an old security plugin. If assistants that search the live web never mention you, and your competitors with similar sites do fine, this is the first thing to rule out.
No structured data. Your site loads fine for people, and to a machine it is a wall of unlabeled text. Nothing says “this is a gym, these are the hours, these are the classes” in a form an assistant can trust. You exist, faintly, and never make the shortlist.
Thin or stale information. The assistant can read you, and what it finds is a homepage with a slogan, a photo gallery, and a timetable from last year. There is nothing specific enough to match against a real question. Beginner friendly? Early opening? Small group strength? If it is unsaid, it is unknown.
No third party mentions. Your own site is in order, and the wider web is silent about you. No directory entries, no local press, no reviews with substance. Assistants lean on outside confirmation before they recommend, so a gym only its own website vouches for stays off the list.
Most gyms have two or three of these at once. The order to fix them: access first, structure second, content third, mentions as the long game.
One more pattern worth naming: the wrong version of you. The assistant describes your old location, quotes hours you changed last year, or attributes classes you stopped running. This is stale data winning over fresh data, and it usually means the outdated sources about you are easier to read than your own website. The fix is the same structure and consistency work, applied with extra attention to the places still carrying the old facts.
Common surprises
Run this test with other owners and the same reactions come up. Someone discovers their site blocks AI crawlers because of a setting nobody remembers choosing. Someone is described entirely from a directory listing they never claimed, wrong prices included. And someone finds a competitor being recommended on the strength of a detailed FAQ page, for questions the competitor answers and they merely could. Every one of those surprises is an action item wearing a disguise.
Your next five minutes
You now know more about your AI visibility than most of the industry. Two ways to go deeper.
The instant route: run the free AI visibility check. It reads your website the way an AI crawler does and scores the ten technical signals, including crawler access and structured data, in seconds. Think of it as the machine half of the test you just ran by hand.
The full route: request a full audit. We put a long list of real questions to the main assistants, score how often you come up against your actual competitors, and hand you a prioritised fix list. It is the difference between a pulse check and a diagnosis, and it is how the Visibility Audit starts every engagement.
Five minutes today tells you where you stand. What you do next decides where you stand in three months.