Every AI visibility vendor tells customers to let AI crawlers in, publish an llms.txt and mark up who they are. We ran the same checks on the homepages of 18 of them, ourselves included.
sameAs links, against 27% of the top 1,000.| Vendor | AI crawlers blocked | llms.txt | JSON-LD | Org + sameAs | Words in HTML |
|---|---|---|---|---|---|
| Profound | None blocked | None | Yes | Yes | 2,237 |
| Writesonic | None blocked | Valid | Yes | Yes | 2,336 |
| LLMrefs | None blocked | Valid | Yes | Yes | 1,381 |
| AIclicks | None blocked | Has errors | Yes | Yes | 2,835 |
| Rankscale | None blocked | Has errors | Yes | Yes | 3,929 |
| Peec AI | None blocked | None | No | No | 3,758 |
| Otterly.AI | None blocked | Has errors | Yes | Yes | 2,972 |
| Bluefish | None blocked | None | No | No | 1,359 |
| AthenaHQ | None blocked | Valid | No | No | 2,500 |
| Evertune | None blocked | Valid | Yes | Yes | 1,291 |
| Goodie | None blocked | Valid | Yes | Yes | 1,908 |
| Knowatoa | None blocked | Has errors | Yes | Yes | 1,519 |
| Gauge | None blocked | Valid | Yes | Yes | 1,753 |
| Scrunch | None blocked | Has errors | No | No | 5,220 |
| Brandlight | None blocked | Valid | Yes | No | 2,867 |
| Hall | None blocked | None | No | No | 204 |
| Am I On AI | None blocked | None | No | No | 969 |
| Mentioned (us) | None blocked | Valid | Yes | Yes | 2,869 |
“JSON-LD” and “words” describe the homepage HTML as the server sent it, before any JavaScript ran, which is how AI crawlers read it according to Vercel’s crawler analysis. A homepage that builds its content in the browser can look complete to a visitor and show few words here.
It says the category takes crawler access seriously: every vendor is open to every AI crawler, which only about four in five of the top 10,000 websites are (our crawler study). On llms.txt and structured data, the vendors are ahead of the web at large but far from unanimous.
It doesn’t say who builds a better product. None of these checks measures data quality, model coverage or price, and none is proven to make an assistant recommend anyone: Google says no special files or markup are needed for its AI features, and no major assistant has said it reads llms.txt. They are cheap hygiene, and the vendors’ own homepages are a fair place to see who practises it. For what the tools cost and cover, see AI visibility tools compared and our study of the vendors’ search footprint.
The 18 vendors are the ones in our October 2026 market study: companies that sell AI answer visibility tracking as a main product, plus Writesonic, a writing platform with GEO features. We are one of them. On October 28, 2026 we fetched each homepage, /robots.txt and /llms.txt once, with the same code and parsers as our crawler, llms.txt and structured data studies, and compared them with those studies’ top-1,000 figures. “AI crawlers blocked” means disallowed from the homepage path by robots.txt. Homepages change often; this is one day’s reading, and any vendor that fixes something after it is welcome to tell us.
You are welcome to quote and chart these figures with a link back. Suggested citation: Mentioned, “Do AI Visibility Tools Follow Their Own Advice? 18 Vendors Checked,” October 28, 2026, https://mentioned.at/learn/ai-visibility-vendors-ai-readiness. Data and text are published under CC BY 4.0.
None of the 18 we checked in October 2026 did. Each robots.txt allowed OpenAI’s, Anthropic’s, Perplexity’s and Google’s crawlers to fetch the homepage.
13 of the 18 vendors we checked, 8 of them without structural errors. Across the top 1,000 websites the share was 22%.