Same questions, three weeks apart: ChatGPT, Gemini and Perplexity went from naming them 6% of the time to 19%.

1 · What we measured

21 questions in plain customer language (urgent needs, everyday situations, specific services, and “who is best” comparisons), in and around the cities the business serves. None of them names the business. Each question ran 3 times on each assistant, because AI answers change from one ask to the next, through each vendor’s API with web search on. We counted an answer only if it named the business.

Our audit covers ChatGPT, Claude, Gemini, Google AI Overviews, and Perplexity. To save cost, the October re-check left Claude out. In September, Claude named the business 0 times in 63 answers.

2 · What we changed in between

• We added a dedicated page for each city it serves beyond its home city, each written for that city, during September.
• We restored its real service facts on every page: its hours, what it does and does not do, and how it quotes a price.
• We completed its Google Business Profile with the services and attributes it was missing.
• We added one plain sentence saying who it is, used word for word across the site, and four articles answering questions its customers ask.

3 · The results

Share of answers that named the business, Sep 10, 2026 vs Oct 2, 2026 (63 answers per assistant on each date):

• ChatGPT: 9 of 63 (14%) to 17 of 63 (27%)
• Gemini: 2 of 63 (3%) to 13 of 63 (21%)
• Perplexity: 0 of 63 (0%) to 6 of 63 (10%)
• All three together: 11 of 189 (6%) to 36 of 189 (19%)
• Google AI Overviews, the same 20 questions asked once each (one could not be measured in October): named the business in 0 of 14 overviews shown in September, and in 4 of 12 in October
• Links to its own website: ChatGPT linked to the business’s site in 8 of 63 answers in September and 17 of 63 in October

For the three assistants together, the rise is larger than the run-to-run noise we see between repeat runs. For ChatGPT alone it is not yet clear of that noise. The mentions are also concentrated: in October, 9 of the 21 questions produced every mention, and the rest never named it.

4 · Where it showed up

Most of the rise came in the cities where the business had a new dedicated page, and in its home city. Sunnyvale, whose page went up just before the October run, had not moved yet:

• Los Altos: 1 of 18 to 11 of 18 (city page added in September)
• Menlo Park: 1 of 36 to 7 of 36 (city page added in September)
• Atherton: 1 of 9 to 4 of 9 (city page added in September)
• Palo Alto: 6 of 63 to 9 of 63 (city page added in September)
• Its home city: 1 of 9 to 5 of 9 (covered by the homepage)
• Sunnyvale: 0 of 18 to 0 of 18 (city page added just before the October run)
• Cupertino and Stanford: 1 of 36 to 0 of 36 (outside its service area, not a clean comparison)

Some answers began repeating facts from the site itself, such as its hours. Some answers that named it also repeated stale or wrong details, which we are working to correct.

5 · What this does and does not prove

It shows a measured rise on the same questions, concentrated where we made changes. It does not prove which change caused it: AI assistants also refresh what they know on their own schedule, one assistant (Perplexity) moved to a newer model between the two dates, and each date is one round of runs. It covers AI answers only, not Google Maps rankings, and says nothing yet about calls or revenue. AI answers vary by person and session, so these are measured rates, not a guaranteed position. This is one client, in a different industry from medspas, so your results will differ. We re-check on October 13, 2026 and will update this page.