We test local businesses across ChatGPT, Google AI Overviews, and Gemini regularly, and a pattern keeps showing up: the businesses that don't get mentioned are rarely the worst ones. Great, well-reviewed, genuinely good businesses across St. Petersburg get skipped over in AI answers constantly, and it's almost never because a competitor is better. It's because the AI engine never had enough to go on.
Search has changed faster than most local businesses have. A customer asking "best acupuncturist in St. Pete" or "who does custom cakes near Grand Central District" increasingly gets a synthesized answer instead of a list of links to click through. Whether a business shows up in that answer depends on a different set of signals than traditional ranking ever did, and most St. Petersburg small businesses haven't touched any of them yet.
The three gaps we see most often, locally
Across audits of St. Petersburg businesses, the same three problems come up again and again, usually in combination rather than isolation.
1. Google Business Profiles that are technically claimed, functionally empty
Almost every business we look at has claimed their Google Business Profile. Far fewer have filled it out completely: services described with actual language instead of one-word labels, current photos, an active posting habit, a correct category. AI Overviews in particular lean heavily on this data. A half-finished profile isn't broken, it's just quiet, and quiet doesn't get recommended.
2. No content built for the specific thing customers actually ask about
A general services page listing everything a business does reads fine to a person. AI engines match much better against a page built around one specific thing, a specific condition treated, a specific service performed, a specific specialty. A wellness studio with a single "Services" page competes for very little. The same studio with a dedicated page on the one thing it's actually known for becomes citable for exactly that.
3. Thin entity authority beyond the business's own website
Schema and a clean website tell AI engines what a business claims about itself. Independent directories, consistent citations, and a steady rhythm of recent reviews tell AI engines that claim is actually true. Most St. Petersburg small businesses have a website and maybe one or two directory listings set up years ago and never revisited. That's a thin foundation for the kind of corroboration AI engines are increasingly built to look for.
Being a good business isn't enough anymore. Being legible to the systems doing the recommending is the part most owners haven't gotten to yet.
Why this matters more here than it might elsewhere
St. Petersburg's small business scene is dense and genuinely strong, restaurants, wellness practices, boutiques, home service providers, competing in a relatively compact geographic area. That density cuts both ways. It means there's real customer demand to be found, and it also means a lot of similarly good businesses are all fighting for the same handful of AI-generated recommendations. The businesses that get specific, structured, and corroborated first are the ones that start showing up consistently, while equally good competitors stay invisible simply because nobody told the AI engines they exist.
Free download
Start with the foundation, not the theory.
The free 27-point checklist covers everything that affects whether AI engines can find, parse, and trust a Google Business Profile. No credit card, just the list.
Get the free checklist →Where to start
Foundation first, always. A complete Google Business Profile and a basic review request rhythm are the fastest, cheapest fixes and the ones most likely to move the needle in weeks rather than months. After that, one dedicated page for the single thing a business is actually known for does more work than a dozen generic blog posts. Entity authority, the directories, the citations, the sameAs links tying everything together, is the slower layer that compounds over months, but it's what turns occasional AI mentions into consistent ones.
None of this requires abandoning what already works. It requires treating AI visibility as its own project, not an afterthought bolted onto a website that was built for a different kind of search entirely.