AI readiness

How Site Health checks whether AI assistants can reach your site, trust it and quote it, and how to fix what it finds.

AI readiness is a group of checks in every Site Health audit. Whether AI assistants like ChatGPT, Claude and Perplexity can reach your pages, trust them and quote them. An AI assistant can only cite a page it can reach and read, so these come before anything you write.

Note

An audit costs up to about $0.33 on your keys. Site Health can crawl 10,000 a month pages per website, across all its audits.

Run an audit

  1. 1Open Site Health in the sidebar.
  2. 2Click Start technical audit, check the cost note, then click Start Audit. A large site can take a few hours, and you can leave the page while it runs.
  3. 3When it finishes, the Audits list shows AI readiness: N of M checks passed for the latest audit. Click it to open the report at that group.
  4. 4Click any check to see what it means, why it matters, how to fix it (with notes for WordPress and Ghost) and the pages it found. Some checks have code to copy.
  5. 5Once you've dealt with a check, click Mark fixed to keep track. Your next audit shows whether it passes.

What it checks

Access: whether AI assistants' bots are allowed in and can read your pages.

CheckLevelWhat it means
Google blocked in robots.txtCriticalYour robots.txt stops Googlebot from crawling some of your key pages.
ChatGPT search blocked in robots.txtWarningChatGPT's search bots can't reach some of your key pages.
Claude blocked in robots.txtWarningClaude's search bots can't reach some of your key pages.
Perplexity blocked in robots.txtWarningPerplexity's search bots can't reach some of your key pages.
Bing blocked in robots.txtWarningYour robots.txt stops bingbot from crawling some of your key pages.
Apple blocked in robots.txtNoticeYour robots.txt stops Applebot from crawling some of your key pages.
AI bots may be refused by your serverNoticeYour server turned away a request that identified as an AI search bot.
Check Cloudflare's AI crawler settingNoticeYour site runs behind Cloudflare, which can block AI crawlers before they reach you.
Content only appears with JavaScriptWarningOn pages built this way, text, the main heading or structured data only appear after JavaScript runs.
Pages that block AI answersWarningA nosnippet or very low max-snippet setting stops Google quoting the page, in results and in AI answers.
Pages that ask AI tools not to use themNoticeA noai or noimageai tag asks AI tools to leave the page alone. Fine if intended.
Pages AI bots never readNoticeNo AI search bot read these pages in the last 90 days, so AI answers can't cite them.
No llms.txt fileNoticeThe site has no llms.txt, a short guide to your site written for AI tools.
llms.txt isn't in the expected formatNoticeYour llms.txt has no title line or lists no links.

Trust: whether your site says who is behind it and how current it is.

CheckLevelWhat it means
No Organization schemaWarningThe homepage doesn't say, in structured data, who runs the site.
Organization schema is missing detailsNoticeYour Organization structured data has no logo or no links to your official profiles.
Articles with no visible dateNoticeReaders and AI assistants can't tell how current the article is.
Articles not updated in over a yearNoticeThe article's updated date is more than 12 months old.

Answer shape: whether your articles are easy to quote.

CheckLevelWhat it means
Articles that don't answer firstNoticeThe opening doesn't give a short, direct answer before the detail.
No question-led headingsNoticeNone of the article's headings are phrased the way people ask.
Long articles with no lists or tablesNoticeLong articles written only as paragraphs are hard to scan and to quote.
Statistics without sourcesNoticeThe article quotes figures but never says where they come from.

Fixing the new checks

  • llms.txt: open the check and copy the starter file. It's filled in from your Brand knowledge and your published articles. Edit it, save it as llms.txt at the root of your site, and open yoursite.com/llms.txt to check it shows as plain text. No major AI assistant has confirmed it reads llms.txt, so treat it as cheap insurance, not a ranking factor.
  • Articles not updated in over a year: refresh the facts first (figures, prices, names, anything new), then the date. Never change the date alone. Improve on each page opens it in Page Audit.
  • Pages AI bots never read: link to them from pages AI bots already read, such as your homepage, menus and popular articles, and check the access checks above. Articles we published for you are listed first.

Pages AI bots never read needs AI Crawlers

This check compares the pages in your audit with the pages AI search bots read, so it needs Crawl Log set up and about 30 days of visits. Until then it shows Not checked with the reason, and a link to set it up.

When a check says Not checked

  • Added after this audit ran: the check is newer than the audit. Your next audit runs it.
  • Still running: AI access tests run for a few minutes after the crawl. The report updates on its own.
  • We couldn't read your robots.txt or llms.txt: your server didn't answer that request. Run another audit later.
  • Needs a full-site audit: a URL check reads only the pages you list, so checks about the whole site are left out.