Why ChatGPT Barely Recommends Any Local Businesses

Fake CAPTCHAs And Fake Trust

Ars Technica reported on September 11 that ClickFix attacks are going viral across PCs and Macs. The setup is simple. You land on a page, it shows you a fake CAPTCHA, and instead of clicking a box, you're told to copy a command and paste it into your terminal to "verify you're human." People do it. Then malware runs.

What makes ClickFix work isn't a technical exploit. It's trust. The page looks familiar enough, the instructions feel official enough, and there's no human on the other end checking whether any of it is real.

That's basically the same problem AI engines face when they decide which local business to recommend. Nobody at ChatGPT is calling your shop to confirm you're still open. There's no human verifying your hours, your address, or whether your reviews are real. The system has to infer trust from signals, the same way a victim infers trust from a CAPTCHA that looks legitimate.

The difference is scale. ChatGPT isn't making one trust decision. It's making millions, across every city, every category, every search. When there's no human checking the work, the system leans hard on structured signals it can verify automatically. Which raises the actual question small-business owners should be asking: what are those signals, and does your business even have them.

The 1.2% Problem

SOCi ran the numbers in its 2026 Local Visibility Index and found something that should worry any business owner leaning on Google rankings alone. Only 1.2% of businesses got recommended when people asked ChatGPT for a local suggestion. Perplexity did a little better at 7.4%. Google's own local pack, the map and three listings you're used to seeing, surfaced businesses 35.9% of the time.

That gap is not a rounding error. It means the overwhelming majority of businesses that show up fine on Google search are simply invisible when someone asks an AI assistant the same question.

Here's the part that trips people up. Ranking well on Google does not transfer to ChatGPT or Perplexity the way most business owners assume it will. Some 2026 studies found only about 38% overlap between who ranks on Google and who gets cited by AI. Different systems, different inputs, different math.

Google's local pack pulls heavily from your Business Profile and proximity to the searcher. ChatGPT is drawing from a different set of data sources entirely, pulling citations, structured data, and third-party listings rather than matching a map pin to a zip code. If your visibility strategy stops at Google, you are optimizing for one gatekeeper while a second one, increasingly used by consumers, never sees you at all.

What AI Engines Actually Check

GeoHero's July research on this got specific. ChatGPT leans on three things when it decides who to mention: a complete Google Business Profile, recent review volume, and LocalBusiness schema sitting in your website's code. Miss any of those, and you're not being punished exactly. You're just not visible to the system doing the recommending.

The Foursquare piece surprised a lot of people who assumed Yelp was carrying most of the weight. Pleiades Consultancy's April analysis found Foursquare supplies roughly 70% of the local business data ChatGPT actually draws from. If your listing there is thin, outdated, or missing entirely, you're cutting off the pipe that feeds most of what the model knows about you.

NAP consistency, your name, address, and phone number matching exactly across every directory, matters more here than it ever did for Google ranking. AI engines cross-reference these listings to decide whether a business is real and current. A mismatched suite number or an old phone listing on Yelp is enough to introduce doubt.

Star ratings matter less than people think. Uberall's September study of 120,000 mentions found review volume predicts whether AI cites you far better than your average rating does. A 4.2-star business with fifty recent reviews beats a 4.9-star business with three reviews from two years ago.

Fixing Your Visibility This Week

None of this requires hiring an agency or waiting on a developer for three months. Start with your NAP data. Pull up your Google Business Profile, your Yelp listing, and your Foursquare listing side by side. Check the name, address, and phone number character by character. A "Suite 200" on one and "Ste 200" on another sounds trivial. It is not trivial to a system cross-referencing thousands of listings to decide if you are real.

Next, get LocalBusiness schema onto your website. If you are not technical, paste your homepage code into ChatGPT or Claude and ask it to identify what schema is missing and generate the markup. Your developer or a WordPress plugin can add it from there.

Reviews need to keep coming, not just exist. Uberall's data pointed to volume over rating, so set up a simple system, a text after every job, an email after every purchase, asking for a review. Recent and steady beats old and perfect.

Then check your Foursquare listing specifically, given how much weight Pleiades found it carries. Most businesses have never touched it.

BrightLocal's Local AI Visibility Tracker, launched September 10, lets you see where you actually show up across ChatGPT, Perplexity, and Google AI Overviews right now, instead of guessing.