Is your dental practice ready for ChatGPT? we measured 385 Nordic practices

We measured 385 dental practices in Denmark, Norway and Sweden. The typical practice can be read by AI, but one in three Danish practices lacks structured data. Here are the figures and the five priority improvements.

See whether ChatGPT mentions your business — free test

Are dental clinics’ websites ready for ChatGPT and AI search? Short answer: Most are. We measured 385 dental practices in Denmark, Norway and Sweden, and the typical practice scores 85 out of 100 when an AI robot scans the homepage. What is most often missing is machine-readable information about the clinic: one in three Danish dental clinics has no structured data, and nearly one in four lacks a meta description. These are minor corrections, and they can be made in an afternoon.

Technology is rarely what keeps a dental practice out of the AI results. When a patient asks ChatGPT, Perplexity or Google’s AI for a dentist in their town, the AI must first be able to read the clinic’s website. It can do this for the vast majority of clinics. But it also needs to be able to work out what the clinic is called, where it is located, when it is open, and what treatments it offers. This is where some clinics make it harder for the machine than it needs to be.

This article sets out what we found, how Danish dentists compare with their Norwegian and Swedish counterparts, and what specific changes you can make straight away. The figures are our own, and we explain at the bottom how they were collected and what we have excluded.

Can AI read a dental clinic’s website?

Yes, in most cases. We retrieved the homepage of each clinic in the same way as GPTBot, ClaudeBot and PerplexityBot do: without running JavaScript. We then measured six technical indicators that determine whether an AI bot can read and understand the page. The median score was 85 out of 100 in all three countries.

Signal	Denmark (n=133)	Norway (n=116)	Sweden (n=136)
Median score (0–100)	85	85	85
No structured data	33%	28%	26%
Without a meta description	23%	20%	11%
Without an H1 heading in raw HTML	15%	13%	12%
Less than 150 words of readable text	8%	3%	4%
With llms.txt	10%	21%	18%
Blocks an AI robot in robots.txt	0%	2%	1%

Two things stand out. Firstly, almost no dental practices block AI robots. This is good, as a blocked robot cannot use the site as a source at all. Secondly, dentists have a higher median than the solicitors we assessed in July. There, the median was 70, and half of the solicitors’ websites lacked structured data. You can read that analysis in the article on solicitors’ AI visibility.

What do Danish dental practices most commonly lack?

The information the machine needs to understand the clinic. Among the 133 Danish clinics, we found:

• 44 clinics (33 per cent) have no structured data. There is no machine-readable description of who the clinic is and where it is located.
• 31 clinics (23 per cent) have no meta description. This is the short text that appears in search results, and is often the first thing an AI engine reads about the page.
• 20 clinics (15 per cent) have no H1 heading when the page is viewed without JavaScript. The heading may appear large to humans but still be invisible to the machine.
• 11 clinics (8 per cent) have less than 150 words of readable text without JavaScript. To an AI robot, the homepage is virtually empty.

24 of the Danish clinics (18 per cent) lack both structured data and a meta description. These are clinics where the AI has to guess most of the information from the body text.

Why is structured data important for a dental practice?

Because it removes the guesswork surrounding the information patients are looking for. Structured data is a small block of code on the page that says the same thing as the text, but in a format that machines can read unambiguously. There’s a specific type designed for dentists – ‘schema.org/Dentist’ – with fields for name, address, telephone number, opening hours and services.

Let’s be clear about what it does. Google states in its guide to AI search that structured data is not a requirement for inclusion in AI Overviews or AI Mode. However, Google does use structured data to display local businesses’ details, such as address and opening hours, and this is precisely the information a patient is looking for. If the clinic has several branches or an address that resembles that of another clinic, this is where confusion arises.

The most important thing is that the block matches what is stated on the page. If one telephone number is given in the text and another in the code, you’ve made things worse, not better.

How do Danish dentists compare with their counterparts in Norway and Sweden?

Slightly worse on most metrics. Swedish clinics are best at meta-descriptions: only 11 per cent are missing one, compared with 23 per cent in Denmark. Both Norwegian and Swedish clinics are more likely to have a llms.txt, whilst Danish clinics are more likely to have a virtually empty homepage for the AI bots (8 per cent compared with 3–4 per cent).

The differences aren’t huge, and the median is the same. But they show that there isn’t any particular Danish problem that can’t be solved. It’s a matter of the same few corrections, and neighbouring countries have simply made them a little more often.

Is a sound technical setup enough for ChatGPT to recommend my clinic?

No, but it’s the baseline. The technical aspects determine whether the AI can read your website and use it as a source. It does not, on its own, determine whether the AI will recommend you over the neighbouring clinic. It also depends on your reviews, what others write about you, and how clearly your website answers the questions patients ask.

In our analysis of 1,912 Nordic companies across 15 sectors, companies with a technical score of 85 were used as a source in the AI responses in 54.5 per cent of cases, compared with 24–39 per cent for companies with the lowest scores. This is a correlation, not proof that the score alone is the cause. The figures and methodology can be found on the data page about technical scores. We’ve summarised the other factors that determine whether AI mentions you in the article on what determines whether AI mentions your business.

If your clinic already scores 85, the next step is not more technical work. It’s the content: are the treatments, prices, opening hours and emergency appointment times listed on the page in plain text so that the AI can quote them?

Does a dental clinic need a llms.txt?

That’s not the first thing you should do. 90 per cent of Danish clinics don’t have a llms.txt, and that’s not a disaster. Google states that Google Search does not use llms.txt. In our own survey of 1,912 businesses, those with the file were used as a source in 43.6 per cent of cases, compared with 43.3 per cent without it. The file does no harm, but structured data, a meta description and a readable front page take precedence.

What should a dental practice fix first?

Start with what provides the most accurate information to the machine with the least effort:

1. Check what the AI robots see. Run the homepage through a free test that reads it without JavaScript. If there are fewer than 150 words, that’s what you should fix first.
2. Add structured data to the homepage. Use the ‘Dentist’ type with the clinic’s name, address, telephone number and opening hours, exactly as they appear in the text.
3. Write a meta description. One sentence containing the clinic’s name, town and the main treatments offered. This appears in the search results and is read by AI engines.
4. Make the heading a proper H1. It should state who you are and where you’re located, and it must be an H1 in the code, not just large text.
5. Include answers to patients’ questions on the page. Appointment availability, prices for standard treatments, and whether you’re accepting new patients. Anything not included there cannot be cited by the AI.

If your clinic uses WordPress, points 2 and 3 can be sorted out with a plugin. If you have a web agency solution, it’s a small task for the agency. If you want it done without having to think about it, that’s exactly what Geoa does for dental clinics: the GEOA AI SEO Employee™ finds the errors and corrects the technical aspects directly on the website.

Frequently asked questions

Are dental clinics’ websites ready for ChatGPT?
Most are. In Geoa’s survey of 385 dental clinics in Denmark, Norway and Sweden, the median score was 85 out of 100 in all three countries, and almost none block the AI bots. What is most often missing is structured data: 33 per cent of Danish clinics have none.

What does structured data mean for a dental practice?
Structured data is a block of code that describes the practice’s name, address, telephone number, opening hours and services in a format that machines can read unambiguously. There is a specific type for dentists: schema.org/Dentist. Google does not require this for AI searches, but uses it to display information about local businesses.

How do Danish dentists compare with their Norwegian and Swedish counterparts?
Slightly worse. 23 per cent of Danish clinics lack a meta description, compared with 11 per cent in Sweden, and 33 per cent lack structured data, compared with 26–28 per cent in Norway and Sweden. The median is 85 in all three countries.

Does a dental clinic need a llms.txt?
That’s not the first thing you should do. Google states that Google Search does not use llms.txt, and in Geoa’s survey of 1,912 Nordic businesses, the file alone made no difference. Structured data, a meta description and a readable homepage take priority.

Is good technical optimisation enough for ChatGPT to recommend my clinic?
No. Technical optimisation determines whether AI can read your website and use it as a source. Whether AI recommends you also depends on reviews, what others write about you, and whether your site clearly answers the questions patients ask, such as about emergency appointments and prices.