AI always names a gym. Claude with live web search named at least one business in all 50 of our fitness questions, with 5.0 per answer, and never hedged. Only restaurants matched that.
The names came from booking and directory sites far more than from gyms' own websites. Directories were 43.3% of sources and gyms' own sites only 14.7%. Lantern, Magicpin, ClassPass and Fresha led, and Fresha was cited mostly through non-English versions (Dutch, Polish, Finnish and others) of its US city pages. Almost one source in five (18.6%) was a spam-style page on an unrelated site, and in San Diego it was 42%. The lists were wide open: 96% of the 233 businesses named appeared in only one answer.
What we asked
50 questions written the way people in the US ask them. The 40 English questions were spread over eight cities, 5 each: San Diego, Las Vegas, Boston, Phoenix, Los Angeles, Atlanta, Dallas and Orlando. They cover ten kinds of request: best gym, a personal trainer for weight loss, a women only gym, the cheapest gym with a pool, yoga for beginners, CrossFit, a physical therapist for a sports injury, a massage spa, Pilates and martial arts for kids. By intent: 20 specific, 16 compare and 4 price. The 10 Spanish questions were set in Seattle and Chicago (5 each), tagged "mixed". Six of them, word for word:
| Question | Language | Intent |
|---|---|---|
| Best gym in San Diego? | English | compare |
| Cheapest gym membership in Las Vegas with a pool? | English | price |
| Best physical therapist in Phoenix for a sports injury? | English | specific |
| Best pilates studio in Orlando? | English | compare |
| Mejor entrenador personal para bajar de peso en Chicago (best personal trainer for weight loss in Chicago) | Spanish | mixed |
| Gimnasio barato con piscina en Seattle (cheap gym with a pool in Seattle) | Spanish | mixed |
How often AI named a gym
All 50 answers named someone and none hedged (study: 93.8% and 6.2%).
| Question type | Questions | Named anyone | Businesses per answer |
|---|---|---|---|
| Compare (best gym, CrossFit, Pilates, trainer) | 16 | 100% | 5.6 |
| Specific (yoga, women only, therapist, massage, kids) | 20 | 100% | 5.2 |
| Price (cheapest gym with a pool) | 4 | 100% | 2.5 |
| Spanish (all types) | 10 | 100% | 4.6 |
Price was the thin spot, as it was across the study. The four "cheapest gym with a pool" questions named 2.5 businesses on average, and the answers said they found little reliable pricing. In Las Vegas, one gym, Las Vegas Athletic Club, was the only one with a clear price in the results. In Dallas, the answer warned that the directory pages it found listed prices in rupees.
About one name in fifteen (6.8%) was a person, mostly physical therapists and trainers. Counting businesses only, 98% of answers named one, 4.8 per answer. Even with names in every answer, Claude was cautious: in at least 36 of the 50 it said there was no single best option or no reliable ranking, and in at least 28 it warned that its sources were marketing or low quality.
Where the answers came from
Claude cited 457 sources for the 50 fitness answers, 9.1 per answer, from 177 different websites, sorted into the same nine types as the main study.
| Type of source | Fitness (457) | All 10 industries (4,587) |
|---|---|---|
| Directories, booking and class sites | 43.3% | 37.3% |
| News and media | 19.3% | 11.0% |
| Off-topic hosts (spam-style pages on unrelated sites) | 18.6% | 15.9% |
| Gyms' own websites and other sites | 14.7% | 29.0% |
| Government sites | 1.8% | 0.9% |
| Social media and forums | 1.1% | 1.6% |
| Review platforms | 0.7% | 3.2% |
| Press releases | 0.7% | 1.2% |
Directories showed up in 42 of 50 answers. These were the sites cited most:
| Site | Citations | Answers | What it is |
|---|---|---|---|
| lantern.llc | 25 | 11 | Local directory with category and booking pages (yoga studios, Pilates, massage) |
| magicpin.com | 22 | 6 | Neighborhood "best fitness centers" pages; 9 of the 22 sat under an /india/ path |
| classpass.com | 21 | 7 | Class booking site with city lists, such as "Best reformer Pilates classes in Phoenix" |
| fresha.com | 21 | 7 | Booking site for massage and trainers; 18 of the 21 were non-English versions of US pages |
| hoodline.com | 20 | 12 | Local news site with city roundups |
| bostonmagazine.com | 11 | 5 | Boston Magazine, cited only for Boston |
Fresha's US city pages were cited in Dutch (5), Polish (3), Finnish (3), Portuguese, Swedish, German and Danish versions, and only 3 times in English. What the answer read was the directory's page for the city and category, in whatever language the search returned.
The spam layer was the third heaviest of the 10 industries, after automotive and real estate: 85 sources from 48 unrelated websites, in 30 answers, and most of the sources in 5. Without naming the sites: a cancer charity's mobile site with "Best Gyms in San Diego", a hardware supplier's subdomain with "Top 10 Best Gyms in San Diego for Every Fitness Goal", a US state government subdomain with "8+ Top-Rated Best Gyms San Diego, CA", and a design portfolio site with a female trainer article.
Review sites were mentioned, not read. 14 answers told people to check Yelp and 21 pointed to Google reviews or Maps, and none cited a page on either.
The gyms named most often
The 50 answers named 233 different businesses and people in 250 mentions, and 95.7% were named only once. These were named most often in our answers:
| Business, as named | Answers (of 50) | Where |
|---|---|---|
| Planet Fitness | 5 | Phoenix, and 4 Spanish answers in Seattle and Chicago |
| Equinox | 4 | Los Angeles, Dallas (pool price), and 2 Spanish answers in Chicago |
| 24 Hour Fitness | 3 | San Diego, Los Angeles, Dallas |
| CrossFit Las Vegas, CrossFit Boston, CrossFit Atlanta | 3 (1 each) | Our matching joined these into one entry, because it strips the asked city from a name. They are three different gyms. |
| LA Fitness | 2 | San Diego, Los Angeles |
The top five as matched took 6.8% of mentions; without the CrossFit merge it would be lower still. Every other spot was decided question by question, from whichever ClassPass list, Lantern category or local roundup matched the words.
Not every name was solid: the San Diego answer listed three gyms found only on spam pages. Our counts keep every name given.
Not a ranking. Counts from one assistant on one day, 21 October 2026, not our ranking or an endorsement. A gym missing here may be excellent, and the lists change from day to day.
City by city
Each city had 5 fitness questions, so these are notes, not rankings. Every answer in every city named someone.
| City (English) | Questions | Per answer | Directory sources | Off-topic sources |
|---|---|---|---|---|
| San Diego | 5 | 5.8 | 31.1% | 42.2% |
| Atlanta | 5 | 5.6 | 67.4% | 4.3% |
| Los Angeles | 5 | 5.6 | 20.0% | 33.3% |
| Boston | 5 | 5.2 | 41.3% | 6.5% |
| Orlando | 5 | 5.2 | 37.0% | 19.6% |
| Las Vegas | 5 | 5.0 | 60.0% | 15.6% |
| Dallas | 5 | 4.6 | 36.2% | 34.0% |
| Phoenix | 5 | 3.8 | 34.8% | 10.9% |
- San Diego had the worst spam layer of any city in this industry: 7 of the 9 sources for "Best gym in San Diego?" were off-topic hosts.
- Boston leaned on its city magazine: Boston Magazine was cited 11 times in 5 answers.
- Phoenix had the shortest lists. For the physical therapist question, 9 of the 10 sources came from a chain called Phoenix Physical Therapy, which the answer said has clinics in Alabama, North Carolina and Virginia, not Arizona. A same-name trap, in a business name rather than a place.
What the Spanish questions showed
| English (40) | Spanish (10) | |
|---|---|---|
| Answers naming at least one business | 100% | 100% |
| Businesses per answer | 5.1 | 4.6 |
| Hedged | 0% | 0% |
| Directory sources | 41.0% | 52.7% |
| Sources with a "best" or "top" style title | 45.9% | 11.0% |
All 10 Spanish questions got Spanish answers with names. 27 of their 91 sources (29.7%) were Spanish pages, above the 23.8% across all Spanish answers in the study, and 7 of the 10 answers cited at least one. Two kinds of Spanish page did the work:
- A chain's own Spanish pages. Planet Fitness has Spanish club pages for Seattle and Chicago. They were cited in 4 Spanish answers, and Planet Fitness was named in all 4.
- Airbnb's Spanish sites. 16 citations in 4 answers, mostly listings of personal trainers on Airbnb. Many were for other places, such as Norwalk and Culver City in California, North Bergen in New Jersey and Washington, DC.
No Spanish fitness answer drifted to another country, but the Chicago trainer answer found mostly other US cities.
Three real answers
Best gym in San Diego?
The search didn't turn up a reliable "best gym" ranking for San Diego. Most results are generic or appear auto-generated, and some contradict each other.
One page includes a placeholder: [Gym Name 1]: The Boutique Fitness Experience.
Cheapest gym membership in Las Vegas with a pool?
I didn't find reliable current pricing for other Las Vegas gyms with pools, so LVAC is the only clearly priced option in these results.
Mejor entrenador personal para bajar de peso en Chicago
Los demás resultados son de Los Ángeles, Miami, Nueva York o Nueva Jersey, o son artículos sin relación directa con entrenadores en Chicago.
What a gym or studio should do
- Complete your profiles on the booking sites that showed up. ClassPass, Fresha, Lantern, Lessons.com and, for CrossFit boxes, CrossFit's own gym pages. Answers quoted what those pages show: class types, neighborhood, prices, first-timer offers.
- Get onto the city lists those sites publish. ClassPass lists such as "Best reformer Pilates classes in Phoenix" were cited directly. Keep your class details and reviews on those sites complete and current.
- Publish one page per class type and location. "Reformer Pilates in Orlando, Florida", "Kids' martial arts in Atlanta, Georgia": schedule, level, price and address. Gyms' own sites were only 14.7% of sources, so a clear page has little competition.
- Put prices on the page. Membership, day pass, intro offer, and whether the pool is at every location. Price questions named 2.5 businesses on average; the answers found little reliable pricing.
- Write the city and state, and your real location. A chain named Phoenix Physical Therapy filled a Phoenix, Arizona answer with clinics in other states. Say where you are on every page.
- If you serve Spanish speakers, write Spanish pages. The one chain with Spanish club pages was named in 4 of 10 Spanish answers. A Spanish page per location, with the city and state, is a source they can use.
- Measure the same questions monthly on ChatGPT, Gemini and Claude, with dated screenshots, and keep Google and Yelp reviews healthy: people check them before they join.
What we'd do for you, and what we don't promise
We do this work for US gyms and studios, remotely from Dubai, with traditional SEO included and ChatGPT and Gemini (Google AI) tracked on every plan. It starts with a free AI check: 10 real member questions and the businesses AI names instead of you. The AI Visibility Audit (30 questions, 4 AI tools, 3 competitors, screenshots and a 90-day plan) is USD 549. Monthly plans are USD 1,249 (Foundation), USD 2,349 (Growth) and USD 4,399 (Authority), month to month, with no lock-in and no setup fee. Details at adjoltz.com/us.
In the first 30 days of any plan, if AI does not name you for at least 3 commercial questions on the list you approved, or you are unhappy for any other reason, we refund 100% of what you paid, in cash, within 7 working days (guarantee rules). We don't promise a top spot on any assistant, rankings, places on any booking site's lists, member numbers, or a US office. We do not have one.
Method and limits. 50 fitness questions from our 500-question US AI visibility study 2026, each sent once, word for word, to Claude with Anthropic's live web search tool on 21 October 2026, between 21:31 and 23:12 UTC. Source types were coded by us with hand-checked lists, so the off-topic share is a floor. Names were matched with simple rules, which joined the three CrossFit gyms above and may make the "named once" share slightly high. One assistant, one day, one run: Pew Research Center found 44% of US adults use ChatGPT and 6% use Claude. We did not check that every named business exists or is open. Every question and its numbers: study data (CSV).