AI Visibility Readiness Checklist
Nine things AI engines need before they can read, quote, and cite your site. Each one is verifiable yourself in minutes, and the free audit measures most of them automatically. Print it, work down it, re-run the audit.
Can AI engines read you at all?
robots.txt does not block the AI crawlers you want citing you
A blocked crawler cannot read your pages, so that engine cannot cite them. Blocking GPTBot is not the same as blocking ChatGPT search, and Google-Extended governs Gemini training, not AI Overviews; the tokens matter individually.
Check it: Open yourdomain.com/robots.txt and read the User-agent lines, or run the free audit, which reports each named AI crawler as allowed or blocked.
Your main content is in the HTML, not injected after load
Most AI fetchers read the raw response. Content that only exists after JavaScript runs is invisible to them, however good it looks in a browser.
Check it: View page source (not inspect) and search for a sentence from your page. The audit measures this as AI Crawlability.
An llms.txt file tells engines what your site is
It is the emerging convention for handing AI systems a clean, quotable summary of who you are and which pages matter, in Markdown they can lift accurately.
Check it: Check yourdomain.com/llms.txt exists and reads like an accurate one-page summary.
Can they quote you confidently?
Every important page answers its question in the first screen
Engines quote passages, not pages. A page that takes four paragraphs to reach its point gives them nothing liftable.
Check it: Read only the first two paragraphs of your key pages: is the answer there, complete, in plain words?
Structured data (JSON-LD) matches what the page says
Schema is how you say, machine-readably, what kind of thing a page is. Mismatched or boilerplate schema reads as noise; accurate schema reads as confirmation.
Check it: Paste a URL into a schema validator and check the type and fields describe the actual page.
Claims carry checkable specifics, not adjectives
Engines preferentially quote concrete, attributable statements: numbers, names, dates. Superlatives without a source are exactly what they skip.
Check it: Pick your homepage's three biggest claims. Could a stranger verify each from the page itself?
One page per question, stable URLs
A page that tries to answer eight questions competes with itself, and URLs that change strand every stored answer that ever cited them.
Check it: For your five most valuable buyer questions: does one distinct, stable URL answer each?
Do they have reasons to pick you?
Third parties say what you say
Engines can cite what OTHERS write about you even when they cannot read you. Consistent third-party descriptions corroborate your own pages; contradictions undercut them.
Check it: Ask an AI assistant what your company does. Compare its sources against your own positioning.
You know which engines cite you today, with dates
Everything above is readiness. Whether it worked is a measurement, and a measurement needs a baseline before any fix can be graded honestly.
Check it: Run the free audit for the structural half; citation sampling over time is what the product itself does.
Readiness is the half you control; whether engines actually cite you is a measurement, not a promise, and completing this list guarantees nothing on its own. It removes the failures that make citation impossible.
Printing this page gives you the paper version; the checkboxes are real.