What does "AI-ready" actually mean? A 13-point checklist

More people now ask ChatGPT, Claude, Perplexity or Google's AI Mode instead of clicking through search results. Those tools can only recommend your site if they can get to it, read it and make sense of it. "AI-ready" is the shorthand for that, but nobody has written an official definition.

A working definition

In practice, a site is AI-ready when AI crawlers (the bots that fetch pages for AI tools) can:

• reach your pages,
• read the content without running JavaScript,
• are allowed to quote it,
• get clear signals about what it is and how current it is,
• and do all of that quickly.

The 13 points below turn that into checks you can actually test. They're grouped into the same four sections as the Jrooky scan results.

First: AI bots aren't one thing

Most AI companies run several bots with different jobs, and blocking one doesn't block the others. OpenAI, for example, uses GPTBot to collect training data, OAI-SearchBot for ChatGPT search, and ChatGPT-User when a person asks ChatGPT to visit a page. Each one is controlled separately, and sites that block OAI-SearchBot won't show up in ChatGPT search answers (OpenAI's crawler docs). Anthropic and Perplexity split their bots the same way.

So blocking training bots is a business decision. Blocking search bots is usually an accident, and it makes you invisible in AI answers.

AI Accessibility: can AI tools get in?

1. robots.txt rules for AI crawlers. Your robots.txt file tells bots which pages they may fetch. Check which AI bots it allows, by name. Jrooky tests your homepage against nine of them, including GPTBot, ClaudeBot and PerplexityBot.

2. Content in the HTML, not just in JavaScript. Vercel found that none of the major AI crawlers render JavaScript (Google's Gemini is the exception, since it uses Googlebot). If your page is an empty shell that fills in after scripts run, most AI crawlers see nothing. Server-render your main text, headings and navigation.

3. Working HTTPS. A valid certificate, and plain http:// addresses that redirect to https://. An expired or misconfigured certificate can make fetches fail outright.

4. An XML sitemap. A sitemap is a list of the pages you want found. Bing says sitemaps are still key to being discovered in AI-powered search.

5. Pages that actually load. Every URL in your sitemap should return a normal page, not an error. In Vercel's data, ChatGPT's crawler hit "not found" pages on about 35% of its fetches, so AI bots already waste a lot of effort on dead links. Keep redirects and your sitemap up to date.

6. llms.txt. A proposed Markdown file at /llms.txt that gives AI tools a short summary of your site and links to your key pages. It's not a formal standard, and major AI companies haven't confirmed they use it. Cheap to add, uncertain payoff. More in What is llms.txt and does your site need one?

Indexing & Citation: can AI tools quote you?

7. No accidental noindex. A noindex tag removes a page from search results, and Google's AI Overviews and AI Mode only link to pages that are indexed and eligible for a snippet, with no extra requirements. Also check that your canonical tag (the "main version of this page" signal) points at your own site.

8. Snippet controls. nosnippet and max-snippet limit how much text Google can show from your page, and Google lists them as the way to limit what appears in its AI features too. The catch: they also shrink your regular search snippets (Google's snippet docs).

Content Signal: is it clear what your page is?

9. Schema markup. Structured data is a standard way to describe a page to machines: this is an organization, this is an article, this is a product. There's no public confirmation that AI assistants rank on it, but it removes guesswork.

10. Fresh, honest dates. An Ahrefs study of about 17 million citations found AI assistants cite newer content than Google's regular results do, with ChatGPT showing the strongest preference. Keep the lastmod dates in your sitemap accurate. Bing says they should reflect real content changes, not the time the sitemap was generated.

Performance: is it fast enough?

11. Core Web Vitals. Google's three user-experience measures: loading (LCP), responsiveness (INP) and layout stability (CLS). The good thresholds are 2.5 seconds, 200 milliseconds and 0.1. They're about site health for people rather than a known AI ranking factor.

12. Server response time (TTFB). How long your server takes to start sending a page. web.dev calls 0.8 seconds or less good, and over 1.8 seconds poor. A slow server means crawlers get through fewer pages.

13. Page size. Googlebot only reads the first 2 MB of an HTML page and ignores the rest. Most pages are far smaller, but huge inline scripts or embedded images can push your real content past the cut-off.

How Jrooky checks it

The free scan runs all 13 checks on your site and marks each one pass or fail, with what it found and how to fix anything that failed.

Run a free scan to see how your site does on all 13.