Is Your Website Invisible to ChatGPT and Gemini?
Nearly half of local searches now go through AI assistants instead of Google — and most small business websites are invisible to them. Here's why, and what to check on your own site.
Written by Dmytro · Founder & lead engineer
More and more people aren't typing a business name into Google anymore. They're asking ChatGPT, Gemini, or Google's AI Mode: "best plumber near Mar Menor," "good real estate agency in San Javier," "where to eat arroz in Murcia centro." The AI gives them two or three names and a short reason why — and then they call one of those two or three.
If your business isn't one of the names it picks, you don't lose a little traffic. You lose the sale entirely, because the person never sees a list of ten options to scroll through. They see three, and they pick one.
This is already happening, not coming
According to BrightLocal's 2026 Local Consumer Review Survey (1,002 US consumers, published March 2026), 45% of consumers used an AI tool to find a local business recommendation in the past year — up from just 6% the year before. Within that group, 31% used ChatGPT specifically and 23% used Google's AI Mode.
That's not a niche behavior anymore. Within a single year it went from "a few early adopters" to "nearly half of people looking for a local business." And the share only grows from here.
Why AI skips businesses that Google ranks fine
Here's the part most business owners don't realize: ranking well in classic Google search and being recommended by an AI assistant are two different things.
Google's search algorithm reads your whole page — headlines, paragraphs, photos, the general gist. An AI assistant answering a question in real time doesn't have time to "read" your site the way a human would. It leans heavily on structured data — a layer of code (schema.org / JSON-LD) embedded in the page that spells things out explicitly: this is a business, this is its name, this is what it does, this is where it is, these are its hours, this is its star rating. Machine-readable facts, not prose the AI has to interpret.
A business can have beautiful photos, great reviews, and genuinely be the best option in town — and still never get mentioned by an AI, simply because nothing on its site tells the AI, in a format it trusts, what the business actually is.
The gap here is real and measurable: Ahrefs analyzed roughly 6 million URLs cited by AI answer engines (published May 2026) and found that 53% of the pages AI tools actually cite carry JSON-LD structured data — about three times the rate found on pages that don't get cited.
We checked — and the gap shows up even among strong, established businesses
To see how this plays out locally, we looked at the actual websites of several well-reviewed, clearly established businesses around the Region of Murcia — the kind of businesses that would reasonably expect to come up when someone asks an AI for a recommendation in their category and area.
A real estate agency near San Javier, with a long track record and solid listings, running on a popular real-estate-specific website platform: no schema.org markup detectable in the page source — just a standard meta description and Open Graph tags built for social sharing, nothing that tells an AI assistant "this is a RealEstateAgent business located here, serving these areas."
A second agency, also well established, on a different real-estate CMS: the same story. A professional-looking site, clearly built by a specialist platform for the industry — and the same missing layer.
A well-regarded, independently run restaurant in the center of Murcia, built on WordPress: a clean page, a clear menu, a solid meta description for search engines — but again, nothing structured enough for an AI assistant to confidently lift as a fact rather than guess at.
None of these are bad websites. They load fine, they look professional, a human visitor has no complaints. That's exactly the point — this isn't about having a good website. It's about having a website that speaks a language AI tools can parse. Plenty of genuinely strong local businesses are running sites that were built for human eyes and classic SEO, from an era before AI assistants were a serious referral channel — and that single missing layer is enough to leave them out of recommendations they'd otherwise deserve to be in.
What this means in practice
This isn't a call to panic or rebuild everything. It's a reminder that "AEO" / "GEO" (Answer Engine Optimization / Generative Engine Optimization) is a distinct discipline from classic SEO, and most existing local-business websites simply weren't built with it in mind — because until recently, it wasn't something that mattered.
If you want to know where your own site stands: look at your page source for application/ld+json. If it's not there, an AI assistant has to guess at facts your site could simply be telling it outright.
Every site we build at Diqoded ships with LocalBusiness structured data baked in from day one, alongside the usual fundamentals — mobile-first, fast-loading, no stock photography, a design that doesn't look like everyone else's. It's not a magic switch that guarantees an AI mentions you tomorrow. But it removes the one technical reason a business that deserves to be recommended quietly isn't.