Why doesn't ChatGPT recommend my business?

Because ChatGPT does not build its recommendations from Google rankings, review counts, or your website — it builds them from third-party sources, and most top-ranked local businesses are simply absent from them. In the only field study of its kind, 70% of businesses ranking on page one of Google Maps were never named when ChatGPT was asked to recommend someone in their own trade and town.

The study: "70% of Top Local Businesses Are Invisible to AI," field data collected and published by Michael Putnam of PresenceScanner (Venice, Florida), August 28, 2026: https://www.presencescanner.ai/ai-search-study-southwest-florida — field data by PresenceScanner; this write-up and analysis by ScriptMasterLabs, published with the author's permission and full credit.

The quotable measurement: 81 local businesses, 8 trades, 5 Southwest Florida towns — every one of them in the top ten Google Maps results for its trade and town — were checked against a single logged-out ChatGPT recommendation request per trade. 70% were never named in ChatGPT's written recommendation; only 30% were named at all. The highest-ranked HVAC company in Venice, with 8,936 Google reviews, was not mentioned anywhere in ChatGPT's answer — not in the recommendation, not in the list above it, not anywhere.

How the study was run

On August 27, 2026, Putnam ran eight Google Maps searches in a logged-out browser — roofer in Venice, house cleaner in North Port, HVAC in Venice, plumber in North Port, painter in Englewood, handyman in Nokomis, pool service in Port Charlotte, landscaper in Venice — and recorded the first ten non-sponsored results for each, in the order Maps listed them (81 businesses in total; house cleaning returned eleven). He then asked ChatGPT, also logged out, one question per trade in the form a real customer would ask: "Who's a good [trade] in [town] FL?" Each question was asked once, not repeatedly.

The naming rule is the part that makes this study worth citing: a business counted as named only if it appeared in ChatGPT's written recommendation — the text a customer actually reads and acts on. ChatGPT also displays a list of businesses above its answer; appearing in that list did not count. That distinction is the whole finding in miniature: the list and the recommendation are two different surfaces, and businesses are losing the one that matters.

The results, trade by trade

Six of the eight trades landed within a few points of each other. Counted by town instead, Venice (30 businesses across three trades) came out 73% not named and North Port (21 across two trades) 71% — the same picture from the other direction. Painting was the best-performing trade at 40% named; HVAC was the worst at 20%. The consistency is the finding: this is not one odd search result, it is the same result eight times across five towns.

Three findings that break the usual assumptions

1. Review count did not protect anyone. The top-ranked Venice HVAC company — 8,936 Google reviews — was absent from ChatGPT's answer entirely. The top North Port plumber, with 891 reviews, was absent. A pool company with 354 reviews, absent. A painting company with 149 reviews, absent. Whatever decides which businesses an AI recommends, a large pile of Google reviews is not sufficient on its own.

2. Having a website made no difference — because the answer is assembled from other sites. Eleven businesses in the sample had no website at all. Four of them were recommended anyway. And when Putnam looked at what ChatGPT actually cited to support its picks, the sources were HomeAdvisor, the Better Business Bureau, Angi, Birdeye, and Expertise — third-party directories and review platforms. Not one business's own website was cited as a source in any of the eight answers. Most local marketing advice optimizes the one surface the AI never quoted: the business's own site.

3. ChatGPT recommends businesses Google does not rank. Across the eight trades, 20 businesses were recommended by ChatGPT that were not in Google's top ten at all — some were not in the Maps results at any position. This is not AI doing Google's job worse. It is AI answering a different question, from different sources, and arriving at a different list.

Two smaller observations from the same fieldwork deserve recording. The star ratings ChatGPT displayed frequently disagreed with the businesses' Google ratings — a Venice air-conditioning company rated 4.8 on Google was shown as 2.5; one HVAC company appeared with three different ratings inside a single answer (4.2 in the list, 5.0 in the written text, 4.9 on Google). And the answers contained wrong-trade inclusions: asked for a handyman in Nokomis, the list included a plumbing company and an appliance repair business; asked for a landscaper in Venice, it returned three paving contractors and a tree service. A real handyman is competing for recommendation space with businesses that do not do the job at all.

What this means: the gap is a measurement problem first

Every input a local business traditionally optimizes — Maps rank, review volume, its own website — was measured in this study, and none of them predicted being named. The variable that did show up in the answers was presence in the third-party sources AI engines actually cite. That makes AI visibility a measurable quantity, not a vibe: either an engine names your business when a customer asks, with sources you can inspect, or it doesn't — and the answer changes independently of your Google position.

That measurement is the work ScriptMasterLabs does. The AEO (AI Engine Optimization) suite at https://scriptmasterlabs.com/aeo exists to audit how a business or product actually appears inside AI answers — named or invisible, cited or uncited, and from which sources. CiteGate, at https://scriptmasterlabs.com/citegate , measures whether AI engines can cite a business at all and what they cite when they do. Putnam's study is the field evidence for why that measurement matters: 70% of the best-ranked businesses in his sample would have scored "invisible" on it, and none of them could see that from their Google dashboard.

What the study does not claim

Putnam states his limits plainly, and they are part of why this data is citable:

• This is 81 businesses in one county, checked in one month (August 2026), against one AI system (ChatGPT). It is a snapshot, not a law of nature.
• The sample is deliberately top-heavy by design: every business in it already ranks on page one of Google Maps. That is not a random sample of local businesses — and it strengthens the finding, because these are the businesses doing best.
• AI answers vary run to run. Each question was asked once, which is what a real customer gets — but a second asking would shift some names.
• The method says nothing about why the systems answer as they do; it reports what they answered.
• The full method is published with the study, the searches are listed, and the whole check takes about two hours to repeat. Putnam invites replication. The raw 81-business dataset is not published and no business is named in the study as an example of failure.

One limit ScriptMasterLabs adds on its own behalf: this is PresenceScanner's fieldwork in Southwest Florida trades, not ScriptMasterLabs' research, and the 70% figure should always travel with its source, sample, and date attached — 81 businesses, 8 trades, August 2026.

Frequently asked questions

Do top-ranked local businesses get recommended by ChatGPT?
Mostly no. In PresenceScanner's August 2026 field study of 81 businesses — all in the top ten Google Maps results for their trade and town — only 30% were named in ChatGPT's written recommendation. 70% were never named.

Why doesn't ChatGPT recommend my business if I rank first on Google?
Because the recommendation is not assembled from Google rankings. In the study, ChatGPT cited third-party sources — HomeAdvisor, the Better Business Bureau, Angi, Birdeye, Expertise — and never cited a business's own website. A business weak or absent in the sources AI engines draw from can rank first on Google and still not be named.

Do Google reviews help a business get recommended by AI?
Not on their own, in this sample. The highest-ranked HVAC company in Venice had 8,936 Google reviews and was not mentioned anywhere in ChatGPT's answer. The top plumber in North Port (891 reviews), a pool company (354), and a painter (149) were also absent.

Does a business need a website to be recommended by ChatGPT?
No. Eleven businesses in the study had no website at all, and four of them were recommended anyway — while businesses with websites and hundreds of reviews were passed over.

How big was the study and who ran it?
81 businesses, 8 trades, 5 Southwest Florida towns (Venice, Nokomis, North Port, Englewood, Port Charlotte), checked August 27–28, 2026 by Michael Putnam of PresenceScanner. It covers one county, one month, and one AI system, and the author states those limits himself. Full study: https://www.presencescanner.ai/ai-search-study-southwest-florida