How a generative engine optimization program made a national fitness chain the most retrievable, parseable and trustworthy source about itself — growing AI-referred sessions 287% year over year.
Our client is a national multi-location fitness chain competing in one of the most search-dependent categories in retail — where a prospective member's decision begins with a question, not a brand. By mid-2025, that question was increasingly being asked of an AI assistant rather than a search engine.
Generative platforms were answering "which gym should I join" with a synthesized shortlist, and the brands named in it were winning the consideration set before a single click. The premise of our program was simple: if AI assembles answers from the sources it can retrieve, parse and trust, the work is to make the brand the most retrievable, parseable and trustworthy source about itself.
Two structural conditions shaped the brief. Nearly nine in ten organic clicks came from branded queries — the brand was mostly being found by people already looking for it. And when AI described the brand, it leaned on complaint aggregators rather than owned content.
Weekly Reddit brand monitoring, cross-referenced quarterly against AI citation data, Search Console question queries and Google People Also Ask, surfaced six recurring member themes — three friction points, three positive signals. Because the same themes appeared in what members write and what AI repeats back, each became a known gap in the answer layer.
Those themes became five plays owned jointly with the client's paid, social, email and PR partners. Working from shared themes meant a single content asset could serve organic, AI retrieval and every other channel at once — which is what let the content program scale.
A roadmap across four page types: core pages rebuilt as the connective thread, FAQ coverage expanded to answer questions on-site rather than cede them to forums, class pages optimized as a non-brand discovery entry point, and priority blog refreshes built against the themes — which became a top driver of AI citations.
ExerciseGym and LocalBusiness schema with location-specific amenity lists, global organization and FAQ schema, standardized Google Business Profile data across every location, and 240 local directory citations across 24 target pages — so generative systems could resolve what each gym is, where it operates and what it offers.
AI visibility became a standing KPI. Semrush AI Brand Performance tracked mentions, share of voice and sentiment across ChatGPT, Google AI Overviews, AI Mode, Gemini and Perplexity; GA4 isolated AI-originated sessions and conversions. Citation data fed straight back into theme identification each quarter.
Each quarter of listening sharpened the next quarter of content, and the content that came out of it is what AI now cites — compounding into a 287% year-over-year rise in AI-referred sessions.
Each row states its own comparison period; before values are derived from the reported year-over-year or quarter-over-quarter change. Sources: GA4, Google Search Console and Semrush AI Brand Performance.
The loop is what compounds. Listen to members, answer on-site, measure what AI retrieves — and the same signal that surfaces the next win surfaces the next problem before you have to guess.
We'll show you what generative platforms say about you today — and where the opportunity is.
Get a Free Audit ↗