2026-11-02 · Last updated 2026-11-02

We checked 200 Australian small-business websites for AI visibility: 46.5% give AI tools no business schema

We read 200 Australian small-business websites the way AI search tools do, and the median score was 35 out of 80. The gap is rarely a blocked bot: only 1% shut out one of the main AI search or answer bots. It's missing detail. 46.5% have no business schema at all, and only 19.5% link their schema to two or more profiles like Google or Facebook.

These are on-site signals that AI search tools can read: robots.txt, structured data, matching contact details, reviews and how pages answer questions. This is not a measurement of what ChatGPT or any assistant says about these businesses. We never asked an assistant anything.

How we checked

We started from 8,073 Australian businesses listed in OpenStreetMap with a website: cafés, restaurants and takeaway, pubs and bars, dentists, doctors and clinics, hair and beauty, and trades. We left out 41 that pointed to a Facebook page or a booking or delivery platform rather than the business's own site, and kept one listing per website.

We drew a random sample with a fixed seed, spread evenly across all eight states and territories and the five types of business, so the totals below are not weighted by how many businesses each state has. We went down that list until we had read 200 websites. That took 243 tries: 43 couldn't be read because the domain no longer resolved, the site timed out or refused us, the page wasn't a working website, or its robots.txt asked checkers like ours to stay out.

On 2 November 2026 we ran each site through the same 6 checks as our free AI visibility checker, scored out of 80:

• AI crawlers: does robots.txt let search and answer bots (Google, Bing, ChatGPT search, Perplexity, Claude) read the home page?
• Business schema: Organization or LocalBusiness structured data with name, address, phone, website and sameAs links to real profiles, per Google's LocalBusiness guide.
• Name, address and phone: written the same way on the home page, the contact page and in the schema.
• Google Business Profile and reviews: a link to the listing, and a rating or reviews on the site.
• Freshness and answers: a visible updated date, and headings that ask a question with a short answer under them.
• llms.txt: optional, and not a proven ranking factor.

We read at most 6 pages per site (robots.txt, home page, llms.txt and up to three linked pages), with our own named user agent, no more than 4 sites at a time and at least 1 second between requests to the same site.

Limits. These are signals on the business's own website. They don't cover what other sites say about a business, which owners report matters more, and they don't tell you whether an assistant names anyone. The freshness check is a rule of thumb. A site that hides details on pages we didn't read may score lower than it deserves.

What we found

The average was 38.2 and the median 35. The best site scored 75.

Very few block AI search bots

Only 2 of the 200 sites (1%) blocked one of the main search and answer bots, so for 99% robots.txt isn't the problem. 8 sites (4%) block training-only bots such as GPTBot while leaving the search bots alone.

Nearly half give AI tools no business details to read

93 sites (46.5%) had no Organization or LocalBusiness structured data on the pages we read, so a machine has to guess the name, address and phone from the page text. Of the 107 that had some, 40 used only the general Organization type rather than a LocalBusiness type. Just 19.5% link their schema to two or more profiles (Google Business Profile, Facebook, Instagram, directories), which is how a machine connects the website to the listings and reviews elsewhere.

The name, address and phone check passed on only 7.5% of sites. One caveat: we compared against the business name as listed in OpenStreetMap, which can differ from how the owner writes it, so this check is harsher here than when you enter your own name in the checker.

Reviews are mostly invisible on the site

47.5% link to their Google Business Profile or a Google Maps listing, but only 15.5% show a rating or reviews on the pages we read, and 48.5% have neither. Only 35.5% have a heading that asks a customer's question and answers it straight away, and 2% show a visible updated date. 28.5% have an llms.txt file; many website plugins now add one automatically.

By type of business and by state

Median score by state: NSW 40, WA 38, VIC 38, ACT 37, QLD 35.5, NT 35, TAS 34.5, SA 30.5. With about 25 sites per state, treat the state gaps as a rough guide, not a ranking.

What to fix first

The checker lists each site's top fixes, biggest gap first. Across the 200 sites, these came up most:

1. Write your name, address and phone the same way everywhere (on the list for 92.5% of sites): home page, contact page, schema and Google Business Profile.
2. Add question headings with a one-sentence answer under each (89%).
3. Add LocalBusiness schema (46.5%): it was the biggest single gap, first on the list, for 93 sites. Our free local business schema generator writes it for you.
4. Link your Google Business Profile (52.5%) from your contact page, and add it to the schema's sameAs.
5. Show your Google rating or a few real reviews (36%).

llms.txt was suggested for 71.5% of sites, but it's optional and unproven, so do it last if at all.

Check your own website, free

The AI visibility checker runs these checks on your site, lists your top fixes and gives you ready questions to ask ChatGPT, Gemini, Perplexity and Claude yourself. For the full list of causes, see why your business isn't showing up on ChatGPT. If you'd rather not do the site work, the AI Visibility Fix does it for you. We do the work; no one can promise an assistant will name you.

Sources

• LeyMish Labs check of public small-business websites, 2 November 2026. Business list from OpenStreetMap, © OpenStreetMap contributors (ODbL).
• Google Search Central, LocalBusiness structured data
• The AI bot list the checker uses: ai-robots-txt/ai.robots.txt (MIT)
• The llms.txt proposal

FAQ