Original research: we asked ChatGPT to recommend tradies in 10 Australian cities, then checked which of the 151 top-rated local businesses it actually named. 84% never came up. Here is what separated the visible 16%.
Your business has hundreds of five-star reviews. You rank in the Google Map Pack. And when a homeowner in your city asks ChatGPT who to call, you do not exist.
That is not a scare line. It is what our data says. In August 2026 we ran a simple experiment: pull the top-rated trade businesses in ten Australian cities from Google's own listings, ask ChatGPT to recommend the best tradies in each of those cities, and count how many of the top-rated businesses the AI actually named.
The answer: 24 out of 151. Only 15.9% of Australia's top-rated trade businesses were named in the AI answers for their own city and trade. The other 84% were invisible.
This post publishes the full findings, the method (so you can replicate it for your own market), and what the visible 16% had in common. The raw summary data is available on request, and the methodology is simple enough that you can run it on your own business in five minutes.
TL;DR
- We tested AI visibility for 151 top-rated Australian trade businesses across 20 city and trade panels (10 cities, 5 trades) in August 2026, using ChatGPT with web search.
- Only 24 of 151 businesses (15.9%) were named when ChatGPT was asked for the best tradies in their own city and trade. 84% never appeared.
- Sydney was the bloodbath: none of the 16 top-rated Sydney plumbers and electricians we tracked were named. Regional cities did far better, with Geelong, Toowoomba and Mackay businesses named at several times the capital-city rate.
- Review quality beat review volume. Businesses ChatGPT named averaged a 4.89 star rating; the invisible ones averaged 4.81 with slightly MORE reviews on average. A 5.0 from 300 reviews beats a 4.6 from 3,000.
- ChatGPT's answers were built almost entirely from Google Business Profile data: names, star ratings, review counts and suburbs, quoted straight from Maps listings.
- Every business we checked had a claimed profile, so claiming your listing is table stakes. The differentiators are rating quality, an accurate profile, and the wider web presence that answer engine optimisation builds.
How we ran the study
The method is deliberately simple, because we want business owners to replicate it.
- We picked 20 city and trade panels. Ten cities (Sydney, Melbourne, Brisbane, Perth, Adelaide, Hobart, Geelong, Mackay, Toowoomba, Darwin) crossed with five trades (plumbers, electricians, roofers, landscapers, pest control), two trades per city.
- We pulled the top-rated businesses for each panel from Google's business listings data: the highest review-count businesses rated above 3.9 stars, up to eight per panel. That produced 151 real Australian trade businesses, every one of them with a claimed Google Business Profile.
- We asked ChatGPT (with web search enabled) one question per panel: "Who are the best [trade]s in [city], Australia? Name specific businesses."
- We matched the answers against the listings. A business counted as visible if ChatGPT named it in the answer for its own city and trade.
One honest caveat: name matching is imperfect. AI answers sometimes shorten business names, so a small number of matches may have been missed. If anything, that means the real visibility number sits near our figure or slightly above it, and nowhere near comfortable.
Finding 1: 84% of top-rated tradies are invisible to AI
Across all 20 panels, ChatGPT produced a confident, specific answer every single time. It always named businesses. They just usually were not the businesses at the top of Google's own ratings.
24 of 151 top-rated businesses (15.9%) appeared in the AI answer for their own market. For the other 127, the enquiry went to someone else, and the business never knew the conversation happened. That is the defining property of AI search loss: there is no page two to be found on. You are either in the answer or you do not exist.
Finding 2: Sydney's top tradies scored zero
The most striking row in the data: not one of the 16 top-rated Sydney plumbers and electricians we tracked was named by ChatGPT. Adelaide and Darwin panels also recorded zero.
Meanwhile the regional and small-city panels performed dramatically better. Geelong businesses were named 5 times across two panels. Toowoomba 4 times. Mackay 3, Hobart 3, Perth 3.
Why? In a market with thousands of competing profiles, ChatGPT's retrieval has enormous choice, and the businesses it surfaces are not necessarily the highest-volume operators. In smaller markets there are fewer credible candidates, so a well-maintained profile has a genuinely better chance of being the answer. For regional service businesses, this is the clearest first-mover signal we have seen: AI visibility is winnable in your market right now in a way it may never be again once competitors notice.
Finding 3: rating quality beats review volume
The named and unnamed groups looked almost identical on paper:
| Named by ChatGPT | Invisible | |
|---|---|---|
| Average star rating | 4.89 | 4.81 |
| Average review count | 311 | 331 |
| Has a website | 92% | 99% |
| Claimed Google profile | 100% | 100% |
Look at that table carefully, because it kills two popular assumptions.
First, the invisible businesses actually had MORE reviews on average. Volume did not save them. What the named group had was a higher average rating: ChatGPT repeatedly surfaced businesses holding a perfect or near-perfect score, often introducing them with lines like "achieved a perfect 5.0-star rating from 372 reviews". A 5.0 from a few hundred reviews beat a 4.6 from thousands, over and over.
Second, everyone has claimed their profile and nearly everyone has a website. Those are table stakes. They get you into the pool; they do not get you picked.
Finding 4: ChatGPT is reading your Google Business Profile out loud
The answers themselves left no doubt about the source. ChatGPT's recommendations arrived formatted with the business name, suburb, star rating and review count, linked to Google Maps searches. It was, in effect, reading Google Business Profile data back to the user with light commentary.
We have written before about how ChatGPT recommends businesses, and this study confirms the local-services version of it: your Google Business Profile is your AI storefront. The categories you chose, the suburb your profile anchors to, your rating, and the words in your reviews are the raw material AI engines compose answers from. If the profile is thin, miscategorised or coasting on a 4.5, the AI has better material to work with elsewhere.
What the visible 16% teach the invisible 84%
Pulling the threads together, the playbook that falls out of this data is concrete:
- Protect the rating, not just the count. Recover unhappy customers before they review. Respond to every review. A steady flow of 5-star reviews with recent dates is the single strongest pattern in the named group.
- Make the profile complete and precise. Right primary category, real services listed, accurate suburb anchoring, photos, and Q&A. The AI quotes what it finds.
- Get the review text working for you. Answers quoted phrases from reviews. When you ask for a review, ask customers to mention the job and the suburb.
- Be present beyond your own site. The engines cross-check. Consistent details across directories, genuine community presence and third-party mentions build the trust that gets a name repeated. This is the heart of answer engine optimisation.
- Test yourself monthly. Ask ChatGPT, Perplexity and Google's AI the questions your customers ask. Track who gets named. Our free AI citation checker automates exactly this.
Frequently asked questions
How was AI visibility measured in this study?
We pulled the top-rated trade businesses (by review count, rated above 3.9) for 20 city and trade combinations from Google's business listings, asked ChatGPT with web search to name the best businesses for each combination, and counted a business as visible if it was named in the answer for its own market. 151 businesses were checked in August 2026.
Why did so few top-rated businesses appear in ChatGPT's answers?
Because AI answers only name a handful of businesses per question, and selection skews toward near-perfect star ratings and complete, precisely categorised Google Business Profiles rather than raw review volume. In large markets like Sydney the candidate pool is enormous, so even excellent businesses are routinely passed over.
Does this mean reviews don't matter for AI visibility?
Reviews matter enormously, but quality beats quantity. The businesses ChatGPT named averaged a 4.89 star rating versus 4.81 for invisible businesses, while the invisible group actually had more reviews on average. Protecting a near-perfect rating and keeping reviews recent matters more than accumulating volume.
Is AI visibility easier to win in regional Australia?
Our data says yes. Sydney, Adelaide and Darwin panels produced zero named businesses from the top-rated group, while Geelong, Toowoomba, Mackay and Hobart businesses were named repeatedly. Smaller markets have fewer credible candidates, so a well-maintained profile and consistent web presence have a real chance of becoming the answer.
How can I check if AI recommends my business?
Ask ChatGPT, Perplexity and Google's AI the questions your customers would ask: best [your trade] in [your city], emergency [trade] near [suburb], and your own business name. Do it monthly and track the trend. Our free AI citation checker at orkkid.com/tools runs this test across engines for you.
Will this be repeated?
Yes. This is the first run of a quarterly panel. We will re-test the same 20 city and trade combinations each quarter and publish movement, so the numbers above become a baseline for how fast AI visibility shifts in Australian local services.
If you want to know where your business stands before the next quarterly run, book a free AI citation audit: we test the real prompts for your trade and suburbs across ChatGPT, Perplexity, Gemini and Google AI Overviews and send you the report. Or start with the AEO service page to see how the visible 16% get built.