Think about the last customer who called you. Before they picked up the phone, they probably asked someone what a job or service like yours should cost. Ten years ago that meant asking a mate, checking a forum, or a making a Google search. In 2026, a growing share of them ask a AI answer engines such as ChatGPT, Perplexity, or Google AI Overviews. They type a question, get a written answer with a few names included, then call one of those names (often without even landing on your website).
Over August and September we tested whether Australian businesses are in those answers and the result for small service businesses is the reason I am writing this.
What we did
We built the Australian AI Visibility Index by selecting 100 Australian businesses across 10 industries: healthcare, real estate, building, manufacturing and legal services on the service side, and ecommerce, supermarkets, major retail, banking and fast-moving consumer goods on the consumer side. We wrote 473 questions a real buyer would ask, covering recommendations, pricing, comparisons, brand-specific queries and problem-based searches. We submitted them to ChatGPT, Perplexity, and Google AI Overviews under consistent conditions and scored 1,459 answers by hand.
The full methodology, including the list of businesses and industries, is published alongside the index.
The number that should worry service businesses
Across all question types, service-industry businesses were named in about half the answers relating to their industry. That is not a disaster. It means the AI systems know these businesses exist.
Then we looked at the pricing questions alone. Across 149 answers to cost-related questions in five service industries, the businesses in our sample were named eight times, that is about 5%.
Pricing questions are where buying decisions happen. "How much does a conveyancer cost in Melbourne?" "What does a bathroom renovation cost in Brisbane?" "What should I pay a physio per session?" A person asking those questions is close to spending money. And in 19 of every 20 answers, the businesses we tracked were not there.
The answer still named someone, it just wasn’t them.
The second problem: named but unlinked
We also checked what happens when a business does get mentioned. Roughly four in ten mentions came with no link to the business's own website.
Where did the AI send the reader instead? To directories, review platforms, forum threads, industry resources and old news articles. Your name appears, and the customer's next click goes to a comparison site with your competitors on it.
One business in the study was assessed across 15 answers and was not correctly identified in any of them. The AI had a business with a similar name, or a fragment of the right details, and filled the gaps with something else. For a small business, that is worse than being absent.
Why this happens to small businesses in particular
The businesses that showed up in pricing answers were not the biggest. The pattern in our data points to how a business is described across the open web, not how much it spends.
Three things separated the visible from the invisible:
Somebody else said it too. AI systems lean on facts repeated from independent sources: industry association listings, local news, review platforms, directories. A business that only describes itself on its own website gives the AI one source. One source is easy to skip.
The price was written down. Businesses with a plain page stating what they charge, or a realistic range, gave the AI something to quote. Businesses with "contact us for a quote" gave it nothing, so it quoted someone else.
The details matched everywhere. Same business name, same address, same service description on the website, Google Business Profile, directories and social profiles. Where those details drifted, the AI lost track of who the business was.
None of this requires a large budget. It requires consistency and a willingness to publish the information a buyer is asking for.
Four things to do this month
Ask the question yourself. Open ChatGPT, Perplexity, and Google AI Overviews. Type the five questions your best customers ask before they buy, including the one about price. Note who is named. Repeat a week later. If you are absent both times, you have your baseline.
Put a price on your website. A range is fine. "Most kitchen renovations we complete fall between X and Y" is quotable. "Every job is different" is not. If you are worried about competitors seeing it, remember they can already get a quote from you in one phone call.
Fix your details in every place they appear. Search your business name and look at every listing that comes back. Correct the name, address, hours, services and category so they match your website exactly. Claim the listings you have not claimed.
Get mentioned by someone who is not you. Join your industry association and complete the member listing. Ask three satisfied customers for a review this week. Send your local paper the story you have been sitting on. These are the sources AI answers are built from, and the work of earning them is what practitioners now call answer engine optimisation.
A caveat, and the point
AI answers change from one run to the next. Our figures are a snapshot of August 2026, and we will run the study again next quarter to see what moves. Treat the 5% as an early signal rather than a fixed verdict.
The fundamentals of search engine optimisation have not gone away. What has changed is that the reward for getting them right is now a place inside the answer, not a position on a page.
But the signal is clear enough to act on. Customers are asking machines what your work should cost, and the machines are answering with someone else's name. The businesses that fix that first will not be the biggest. They will be the ones that made themselves easy to find, easy to quote and easy to link to.