Your practice website probably tells machines more than you think
If you run a medical practice, you may never have added a line of structured data to your website. We looked at 100 small Australian practices, and most of the ones we could read had it anyway. Here is what we found, and what it does and does not mean for you.
How many small practices describe themselves to machines?
We drew 100 practices at random, 50 in metropolitan areas and 50 regional. We could read 78 homepages. 49 of those (63%) carry JSON-LD structured data, the code that tells a search engine or AI tool your name, address and services in a fixed format. 24 (31%) carry none of any kind.
Of the 78:
- 36 (46%) describe the practice as an organisation, and 45 (58%) describe the website.
- 64 (82%) set a canonical address and 55 (71%) write a meta description.
- 44 (56%) set a social sharing image, and 39 (50%) have exactly one main heading.
How did the practice get that markup?
Probably not by choosing it, though we cannot see how any site was built. Website builders and search plugins add this code automatically. That could explain why the share is higher than we found among the largest listed companies. We did not test that, and we cannot see how any site was built.
The catch is what the automatic code cannot say. It does not know which services you offer, who you see, how to book, or whether you bulk-bill. Those facts have to come from you, in the page text and, ideally, in the markup.
Does it matter where the practice is?
A little. The gap between city and regional practices is modest, and each group has only 50 practices, so we would not push it. This is what it looks like:
| Of the readable homepages | Metropolitan (38) | Regional (40) |
|---|---|---|
| JSON-LD structured data | 26 (68%) | 23 (58%) |
| Meta description | 29 (76%) | 26 (65%) |
| Social sharing image | 26 (68%) | 18 (45%) |
| Exactly one main heading | 20 (53%) | 19 (48%) |
The widest gap is the social sharing image. Fewer than half of regional homepages set one, which is the image that appears when a link to the practice is shared in a message or a post.
Are practices telling AI tools what they may read?
Rarely. 69 of the 100 publish a robots.txt file, and 57 of those (83%) point to a sitemap. Fewer than 10 name an AI crawler, and fewer than 10 close the site to one. We do not give exact numbers below 10. That is a decision about training and access, and most practices in this sample have not stated one.
17 publish an llms.txt file. That is a count of files, not of practices that chose to publish one. We opened a sample: some were written by the practice, and at least one was the generic text a popular search plugin produces by default. llms.txt is a proposal, not a standard.
What should you do?
Look at what a machine reads on your homepage, then check that your page text says the specifics: your services, your suburb or town, who you see and how to book. Our free AI Visibility Toolkit at /ai-toolkit checks a homepage in about a minute. If you would rather check it yourself while you are on the page, AISearch Pulse has a free version for Chrome: page checks, each tested against hand-checked web pages (November 2026), a short fix for each one that fails, no licence key, available worldwide. It shows no AI answers, and it is optimised for Australian businesses.
What this study does not say
It says nothing about the quality of any practice or clinician. It reads page code, so markup added later by JavaScript is invisible to us and our structured-data figures are a minimum. We could not read 22 of the 100 homepages, which tells us about our scanner, not about those practices, so every homepage percentage above is out of the 78 we could read. The list comes from open map data, so it covers practices someone has mapped with a website address, which is not every practice. We removed duplicate sites and directory pages and drew at random with a fixed seed. "Metropolitan" means within about 45 km of a state capital's centre. The method, the terms and the limits shared by all our scans are in How AISearch Global runs a technical-setup scan.
Frequently asked questions
Do I need to add structured data to my practice website?
Check first. If your site was built on a common platform you may already have it. Whether it is accurate and complete is a separate question that only someone who knows the practice can answer.
Is a practice listed in a directory already described to AI?
Directories describe you, not your own website. This study looked at practice homepages only.
Should my practice block AI crawlers?
We make no judgement on it. It is a choice about training and access.
Does this say anything about how good a practice is?
No. It describes website setup only.
Where did the list come from?
Open map data, drawn at random. We do not publish the list or name any practice.
Sources
- AISearch Global small-practices snapshot, 22 November 2026 (private evidence held; aggregate counts only published).
- Practice sample drawn from OpenStreetMap data. Contains information from OpenStreetMap, © OpenStreetMap contributors, available under the Open Database Licence (openstreetmap.org/copyright).
- llmstxt.org, "The /llms.txt file" (page opened 22 Nov 2026; presents llms.txt as a proposal).
Viveka Mohan Das
Answer Engine Optimisation (AEO) and GEO Strategist, Published Author & Founder, AISearch Global. More about Viveka
Check your own homepage
The free AI Visibility Toolkit reads your homepage the way a machine does and reports what is there.