AI visibility check
Is ChatGPT recommending you?
Your buyers are asking AI assistants who to hire. If you are not in the answer, a competitor is. Before an assistant can name you, it has to be able to read you. Paste your website and see how ready it is, scored out of 100, with the fixes.
See what a finished one looks like: our own report
How it works
Seventeen checks, four questions, one score.
01
Can the AI assistants get in?
robots.txt, whether the crawlers behind ChatGPT, Claude, Perplexity, Gemini and Copilot are invited there, and whether your server actually lets each one in when it knocks. Then Content Signals and your sitemap.
02
Can an agent read you without a browser?
llms.txt, a Link header that points to it, markdown on request, and whether your words are in the HTML or arrive with JavaScript most crawlers never run.
03
Is there anything worth citing?
Title and description, canonical, Organization structured data, Open Graph, and pages written and marked up as answers to buyer questions.
04
Can it tell who you are?
One address for the site and contact details an agent can read.
What the score means
19 checks across 5 categories, each worth what it moves. The scale grades what an answer engine needs from a company website, not the API plumbing a software product publishes for agents, so a services business can reach 100.
- 90–100
- AI-ready. Assistants can crawl, read, verify and cite the site.
- 75–89
- Citable. Most of what an assistant needs is there; a few gaps still cost citations.
- 50–74
- Readable. Crawlers can read it, but there is little to quote and not enough to verify.
- 25–49
- Crawlable. It can be fetched, and that is about all.
- 0–24
- Invisible to AI. Blocked, empty-handed or both.
The score says whether you can be cited. Whether you are is the citation check we run by hand from the report page.
ArcClaw™ AEO fixes what the report finds, then tracks 50 of your buyers’ questions weekly and publishes the content that changes the answer — from $499/mo.
See ArcClaw™ AEO