Study · October 2026
AI visibility of Berlin businesses 2026
We automatically checked 579 websites of Berlin restaurants, medical practices, law firms, trades and service businesses. The result: almost all are reachable, but most lack the data Google and AI assistants need to understand a business.
Clieni Media · · Data as CSV ↓
Key figures
57.5 %
of websites have no structured data (schema.org). Google and AI assistants have to guess who the business is.
10.4 %
provide address and phone in machine-readable form. Exactly what an answer like “Which dentist in Schöneberg is open?” needs.
87.4 %
have no llms.txt, the file that helps language models grasp a website quickly.
1.4 %
block AI crawlers in robots.txt. So the problem is not blocking, it is missing data.
All checks
Share of the 579 reachable websites that pass the check.
- Encrypted (HTTPS)90.5 %
- Set up for phones84.1 %
- Sitemap present74.6 %
- Meta description set66.1 %
- Exactly one main heading (H1)47.8 %
- Share preview image34.5 %
- Business described as a schema.org type31.1 %
- llms.txt present12.6 %
- Address and phone in schema10.4 %
By industry
Restaurants and cafés
- n
- 123
- Avg. score
- 71/100
- Schema.org
- 27.6 %
- Address + phone
- 11.4 %
- llms.txt
- 11.4 %
- Meta description
- 64.2 %
Medical and dental practices
- n
- 130
- Avg. score
- 73/100
- Schema.org
- 30.8 %
- Address + phone
- 10 %
- llms.txt
- 13.1 %
- Meta description
- 66.2 %
Law firms
- n
- 89
- Avg. score
- 71/100
- Schema.org
- 29.2 %
- Address + phone
- 12.4 %
- llms.txt
- 9 %
- Meta description
- 69.7 %
Trades
- n
- 121
- Avg. score
- 73/100
- Schema.org
- 39.7 %
- Address + phone
- 11.6 %
- llms.txt
- 14.9 %
- Meta description
- 70.2 %
Hairdressers, beauty, bakeries, opticians
- n
- 116
- Avg. score
- 70/100
- Schema.org
- 27.6 %
- Address + phone
- 6.9 %
- llms.txt
- 13.8 %
- Meta description
- 61.2 %
What it means
The tech is mostly fine: nine in ten websites are encrypted, most adapt to phones and respond quickly. What is missing are the details a machine needs to identify a business with certainty: name, type of business, address, phone, opening hours, in a fixed format.
For Google Maps and AI assistants such as ChatGPT, Gemini and Perplexity, that is the difference between being named and being passed over. An assistant asked which law firm or restaurant in a district fits can only recommend what it clearly understands.
Notably, part of the llms.txt files we found are generated automatically by site builders and SEO plugins (at least 24 of 73). The file alone does not create AI visibility; clean, consistent business data does.
Three things every business can do right away
- 01Add business data as a schema.org type (for example Restaurant, Dentist, LegalService) with address, phone and opening hours.
- 02Write name, address and phone the same way everywhere: website, Google Business Profile, directories, social media.
- 03Give every page a clear main heading and its own meta description.
Method
Population: 4,903 Berlin businesses with a website listed in OpenStreetMap (restaurants and cafés, medical and dental practices, law firms, trades, and hairdressers, beauty, bakeries and opticians), retrieved on 6 October 2026. Excluded: pure platform and social media links.
Sample: 661 randomly drawn websites, up to 140 per industry. 579 were reachable and evaluated, 71 did not respond, 11 were not fetched because their robots.txt disallows all crawlers.
Check: home page, robots.txt, sitemap.xml and llms.txt with the free Clieni Media website check, without Google PageSpeed. Only files in the intended format (heading with #) count as llms.txt.
Limits: OpenStreetMap does not list every business, so the sample is not representative of all Berlin companies. Only the home page was checked, automatically. Individual websites are not named.
Citation and data
Clieni Media (2026): AI visibility of Berlin businesses 2026. Sample of 579 websites, as of 6 October 2026. https://clienimedia.de/en/study/ai-visibility-berlin-2026
The results are licensed CC BY 4.0: free to use with attribution and a link. Base data © OpenStreetMap contributors, ODbL.
Free · under a minute
How visible is your website?
18 checks on tech, Google and AI assistants. No sign-up.