Google AI Overviews and your gym's SEO: what actually changes and what doesn't
The short version: Google AI Overviews have little effect on the part of your SEO that brings in members and a lot of effect on the part that used to bring blog traffic. If your local strategy is solid, a well-maintained business profile, reviews, neighbourhood searches, you're considerably more protected than almost any other online business. If your plan was to live off articles like "how many calories does spinning burn", that traffic is dying and it's not coming back.
Now the detail, because the nuance is where the money is.
What AI Overviews are
When you search something on Google, you increasingly get an AI-generated block above the traditional results. It's a summary that answers your question directly, citing a few sources in small text on the right. That's an AI Overview.
The practical consequence: many searches end there. The user reads the summary, gets what they need, and clicks nothing. This is called zero-click search, and the industry figures going around point to a 20–40% drop in CTR for the first organic position when there's an Overview above it. The click you used to earn with a good article now dies in the summary.
This is real and it's growing. But the effect on your gym depends on what type of search we're talking about, and almost everyone lumps both types together.
The good news: the searches that bring you members still work the same
Try it right now. Search "gym near me" or "gym in Chamberí" from your phone. What comes up at the top? The map pack: a map with three business listings, their reviews, their hours, and a directions button. No AI summary.
That makes sense. When someone searches for a gym in their neighbourhood, they don't want a 200-word write-up on what a gym is. They want to see which ones are nearby, which has the best reviews, and which is on their way to work. Google knows this, which is why local transactional searches still show the map pack as the main result. A generated summary would be a worse product.
This makes your gym one of the most protected businesses from the change. A recipe blog competes against AI on its own turf: generic information that a model can summarise without losing anything. You compete on something AI can't generate: you physically exist on a specific street, with equipment, showers, and a schedule. The searches that precede a membership ("gym + neighbourhood", "crossfit near me", "24-hour gym + city") are exactly the ones that still show local results.
If you still don't have the basic local SEO in order, that remains job number one, with or without AI. The local SEO guide for gyms covers the full foundation, and Google Business Profile optimisation is the piece that moves the needle most.
Where it does hit you: informational searches
The blow comes from the other side. Searches like "how much does a gym membership cost per month", "best gym for absolute beginners", or "what to bring on your first day at the gym" used to be blog territory. If you wrote a good article on any of them, some of your traffic came from there.
That traffic shrinks. The user asks, the Overview answers with average prices and three tips, and nobody clicks through. Your article might appear as a cited source, which is fine, but the citation generates far fewer clicks than ranking first used to.
Does that mean informational content is useless? No. It means its function changes. It used to be a traffic machine; now it's an authority and citation machine. And that demands writing it differently.
What to do: be the source the AI cites
Language models cite what's citable. A paragraph of generalities ("the price depends on many factors") is useless to an AI building a response. A specific data point ("in Madrid, the average monthly fee at a neighbourhood gym runs between €35 and €55, boutique gyms between €90 and €150") is exactly what the model wants to reference.
Three practical changes to how you write:
- Specific numbers instead of adjectives. Prices in euros, hours, square metres, number of weekly classes. Every verifiable fact is a citation hook.
- Direct answers up front. If the article answers "how much does it cost?", the figure goes in the first paragraph, not the eighth. Overviews pull fragments; make it easy for them.
- Structured data on your website. LocalBusiness schema with prices, hours, address, and marked-up reviews. It's the format machines read unambiguously, and most gym websites don't have it.
This connects to a broader shift we already covered in the article on Google's AI mode: the search engine is becoming an assistant that answers, and the sources feeding that answer are chosen for clarity and verifiability, not length.
Reinforce what AI can't summarise
Here's your structural advantage. An Overview can summarise what functional training is. It can't summarise how your gym smells at 7pm, whether people say hello when they walk in, or whether the morning class instructor knows everyone's name.
The unsummarisable is where you invest:
- Reviews. Volume, frequency, and responses. They're the content users read after the Overview, right before deciding. Getting reviews systematically is probably the local SEO action with the best effort-to-result ratio there is.
- Real photos. Of your gym, your people, your classes. Not stock. A profile with 40 real photos converts better than one with a logo and two renders.
- Community signals. Events, posts on your profile, mentions in local media and neighbourhood websites. All of that builds the entity "your gym" in Google's knowledge graph, and that entity is what AI retrieves when someone asks about gyms in your area.
Conversational search: when people ask AI directly
There's a new scenario growing fast: someone asks ChatGPT, Gemini, or Google's AI mode "what gym would you recommend in Malasaña for someone just starting out?". The response is a short list, usually two to four names, with one sentence about each.
How do you get onto that list? Nobody has the exact formula, but the observed patterns are consistent: models recommend businesses with a coherent presence across multiple sources, a complete Google profile, a website with clear data, plenty of recent reviews, mentions in directories and local press. In other words, they recommend whoever was already winning at traditional local SEO. The new signal is consistency: if your hours say one thing on Google and something else on your website, you're a worse candidate for an AI that doesn't want to be wrong.
What does change is the language of the questions. People talk to AI like a person: "for beginners", "not too expensive", "with morning classes". Your content should answer those phrasings. A page that explicitly says "beginner-friendly gym in X: what your first week with us looks like" has more chance of feeding that response than a generic homepage. If you compete in a specific niche, SEO for boutique and niche gyms goes deeper on this angle.
Change the metric, not just the tactic
If you're still measuring your SEO by web sessions, you're going to draw wrong conclusions. Impressions without clicks go up, blog traffic goes down, and neither tells you anything about your business.
The metrics that actually matter in 2026:
| Metric | Where to measure | Why it matters now |
|---|---|---|
| Google Business Profile views | Business Profile, performance | Your real "homepage" for local search |
| Profile actions (calls, directions, website clicks) | Business Profile | Direct conversion, immune to zero-click |
| Leads by source | Your CRM or forms | The only thing that pays bills |
| New reviews per month and average rating | Business Profile | Primary signal for users and for AI |
| Brand impressions ("your gym name") | Search Console | Measures whether the rest of your marketing is generating demand |
Notice that blog organic traffic isn't in the table. Track it if you want, but as an authority thermometer, not a business KPI. Real conversion happens through the profile and direct channels, and measuring local SEO return deserves its own system.
What we honestly don't know
The pace. Nobody knows how fast this will roll out or how far it will go. Google experiments, removes Overviews from some searches and adds them to others, and what's true today for "gym near me" could look different in two years. Anyone selling you certainties about the 2028 search engine is selling you smoke.
But look at what's robust under any scenario: an impeccable profile, consistent reviews, a website with clear and structured data, a coherent presence everywhere, and an operation that converts interest into visits. All of that wins in a map pack world, wins in an Overviews world, and wins in a conversational assistant world. Good local SEO was the answer before AI and still is after it; what's changed is the penalty for doing it halfway.
And when that work starts generating enquiries, remember that AI is also on your team: tools like Pilotium use agents to respond to every lead within minutes and take them all the way to the visit, which is the part of the funnel where gyms lose more money than from any Google summary.
This week, run two searches from your phone: "gym + your neighbourhood" and "best gym for beginners in + your city". Look at what comes up in each and where you appear. That snapshot is worth more than any prediction.