AI readiness test: is your site ready for AI assistants?
Checks whether AI assistants such as ChatGPT, Claude, Perplexity and Gemini can reach, read and quote your page.
What does the AI readiness test measure?
The test scores the five conditions an AI assistant needs to use your page as a source. Weights reflect their effect on visibility in answers.
Access
robots.txt is resolved separately for 17 bots, the page is actually requested with 6 bot identities, and noindex, X-Robots-Tag, llms.txt and the sitemap are checked.
Extractability
The main text is extracted from raw HTML without JavaScript; word count, text-to-HTML ratio and JavaScript dependency are measured.
Chunkability
Text is split at H2/H3 boundaries; chunks that start with references like “This” or “Also” and chunks that never name the brand or topic are counted.
Semantics
JSON-LD is parsed; Organization, WebSite, BreadcrumbList, page type, @id links, sameAs and FAQ schema consistency with the visible text are checked.
Citability
Question-style headings, short answers right below them, numeric statements, date and author info, and template repetition across pages are measured.
How is the score calculated?
The total is a weighted average of the layers below. Checks within each layer are scored on their own; a critical issue can cap the total.
| Layer | Weight | Measures |
|---|---|---|
| Access | %25 | Can bots reach the page? |
| Extractability | %20 | Is the text readable without JavaScript? |
| Chunkability | %15 | Does each section make sense on its own? |
| Semantics | %20 | Does structured data say who you are? |
| Citability | %20 | Are there clear, verifiable answers? |
What this test measures — and what it cannot
We analysed the raw HTML without running JavaScript, because most AI crawlers read pages the same way. robots.txt rules were resolved per bot following RFC 9309. Live requests were sent with 6 bot identities (User-Agent imitation; real bots are also IP-verified). What we cannot measure: how often bots actually visit, whether your content is in a model’s training data, or whether an assistant will cite you for a given question.
This report is an automated first assessment. It does not guarantee appearing in AI assistants or Google; it only shows measurable technical and content signals.
Frequently asked questions
What is an AI readiness test?
An AI readiness test is a free analysis that measures whether a web page can be reached, read and quoted by assistants such as ChatGPT, Claude, Perplexity and Gemini. It returns a score out of 100, a grade and prioritised fixes.
Does a high score mean I will appear in ChatGPT?
No. A high score shows your page is technically and editorially ready to be quoted; nobody can guarantee which sources an assistant picks. The score helps you remove the obstacles you control.
Does blocking GPTBot hurt AI search visibility?
GPTBot crawls for model training; to be cited in ChatGPT search you need to allow OAI-SearchBot and ChatGPT-User. That is why the test scores training and search crawlers separately.
How much load does the test put on my site?
An AI readiness scan sends at most 12 requests: the page, robots.txt, llms.txt, the sitemap, the page with 6 bot identities and 2 sample internal pages. Repeat scans within 10 minutes are served from cache.