Does ChatGPT know about Australian businesses?

What we found, and how the sample was built

ChatGPT never named 162 of the 300 Australian businesses we asked about, which is 54% of the sample. Gemini never named 66 of the 137 businesses that returned a complete Gemini record, which is 48%. Counting the two together, 56 of the 137 with complete records from both engines were never named once by either, which is 41%.

How the sample was built matters more than the headline, so here it is in full:

• 300 businesses, 50 in each of six industries: accountants and advisers, cosmetic and skin clinics, dentists, e-commerce and retail, law firms and tradies.
• Five mainland capital cities for the local-service niches.
• Scanned on 30 July 2026, in one week, as a snapshot rather than a trend.
• 1,483 ChatGPT questions and 653 Gemini questions in total, each asked once, with every answer stored word for word.
• The sample frame was the top organic search results for buyer-style queries, plus a set of recognisable national anchor brands in each industry.

That last point is the one that makes the finding mean something. Every business in this sample was already search-visible. These are not obscure businesses nobody can find. They are the ones already winning in classic search, and the engines still did not name more than half of them.

Which Australian industries fare worst and best

Tradies fare worst and e-commerce fares best, and the gap between them is larger than the gap between any two other industries in the study.

Why Australian coverage is thinner than US coverage

We have not measured US coverage, so treat this section as reasoning rather than as a finding. Everything measured on this page is Australian and is in the file. What follows is an explanation, and we would rather label it than let it borrow the authority of the numbers above.

• There is simply more English-language web content about American businesses. These systems learn from and retrieve from what exists, and the Australian corpus is a fraction of the size.
• Our review and directory ecosystems are smaller and more fragmented. Fewer independent pages exist saying that a given Australian business is good at a specific thing, which is exactly the material a recommendation is composed from.
• A lot of Australian local reputation lives where retrieval cannot reach it. Community Facebook groups, closed forums, word of mouth, local radio and print. Real reputation, invisible to a crawler.

Our own data is at least consistent with that explanation. Across all 300 websites, on-site AI readiness averages 74.9 out of 100 while the businesses behind them average 38.3 for visibility, and across the 137 with complete two-engine records the two correlate at Pearson r = 0.116. Australian websites are, on average, in reasonable shape. What is thin is everybody else's pages about them.

What thin coverage means for a business that acts first

Thin coverage is the opportunity, because the bar to be named is lower where fewer competitors are named at all.

Trades is the clearest case in our own file. Sixteen of the 21 trade businesses with complete records were never named by either engine, so on the day we asked, the engines had very few Australian trade businesses to volunteer for most of these questions. In a category that empty, being one of the few with genuine external corroboration is worth far more than the same effort would be worth in e-commerce, where only 2 of 22 went unnamed and the answers already have plenty to choose from.

Two honest limits on that, stated here rather than in a footnote:

• This is a reading of one day. Nothing in our data measures how quickly a thin category fills up, because we did not re-scan the sample.
• Thin does not mean easy. The businesses that were named still had corroboration on pages they did not own. The bar is lower, not absent.

What that work looks like for a business starting from nothing is set out in does AI visibility work for a new business.

Check your own category in two minutes

Ask ChatGPT and Gemini the question your customer would ask, in your city, with no business name in it. That single question tells you more about your category than any benchmark we can publish.

1. Open a fresh chat with memory off in each engine.
2. Ask the buyer question. Note every business named, and in what order.
3. Ask it twice more. The three answers will not be identical, which is itself the most useful thing you will learn.
4. Count how many distinct businesses were named across the three. A short list means a thin category and a real opening.
5. Ask each engine what it knows about your business by name, and check the answer for accuracy.

If nobody in your category is being named at all, that is not a reason to relax. It means the answer is currently being composed from directories and general advice, and the first business with real third-party corroboration will fill that space. Our own figures, each with its denominator, are on the statistics page, and who we are and where we are based is on the about page.