How the AI Visibility Check Works

We crawl what AI crawls

Your homepage, robots.txt, and structured data, read the same way GPTBot or Google-Extended would read them, not the way a human sees the page in a browser.

robots.txt: blocking AI crawlers

User-agent: GPTBot
Disallow: /

User-agent: ClaudeBot
Disallow: /

User-agent: PerplexityBot
Disallow: /

robots.txt: allowing AI crawlers

User-agent: GPTBot
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: PerplexityBot
Allow: /

Sitemap: https://yoursite.co.uk/sitemap.xml

We check specifically for GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and OAI-SearchBot: the crawlers that currently power ChatGPT, Claude, Perplexity, and Google AI Overviews. A rule blocking any of them is a fail, not a warning, because it means that system can't read your site at all.

We score four categories

Every check rolls up into one of four weighted categories. The weights aren't arbitrary: they reflect how much each one actually determines whether an AI system can use your site as a source.

Crawl access5 checks

Structured data5 checks

Answer readiness4 checks

NAP consistency3 checks

Structured data carries the most weight because it's the most direct way to tell an AI system what your business is. A full breakdown of all seventeen checks lives on what we check.

We generate the fix

Every failed check becomes a specific, plain-English fix, and where we have an artifact to generate (schema, llms.txt, FAQ content), it's built from facts confirmed on your own site.

A failed check

schema-localbusiness-missing

{
  "id": "schema-localbusiness-missing",
  "category": "structuredData",
  "status": "fail",
  "weight": 10,
  "evidence": "No JSON-LD of type LegalService/LocalBusiness found on /"
}

The generated fix

Paste-ready LegalService JSON-LD

{
  "@context": "https://schema.org",
  "@type": "LegalService",
  "name": "Example Solicitors",
  "telephone": "TODO_CONFIRM"
}

Notice the TODO_CONFIRM placeholder: we don't have this business's phone number, so we flag it rather than inventing one. Every generated artifact works the same way.

We'll never promise you'll rank in ChatGPT or get cited by AI Overviews. Nobody controls that, and any tool that says otherwise is selling you something. What we can do is fix everything on your end: that's the whole deal.

Questions about the process

How long does a check actually take?

Seconds, not minutes. The engine fetches your homepage, robots.txt, and sitemap once, runs all seventeen checks against that single crawl, and returns a score. Fix generation (the JSON-LD, llms.txt, and FAQ content) runs separately, only once you ask for the full report.

Do you store the content of my pages?

We fetch your public pages to run the audit and generate your fixes. We don't sell, share, or use that content to train anything. The score, the checks, and a content hash are what get stored, so we can tell later whether your site has actually changed.

What if my site fails almost every check?

That's the normal starting point for most sites: AI crawlers are new enough that very few businesses have deliberately optimised for them yet. Every fail comes with a specific fix, not just a lower number, so a bad score is a to-do list, not a verdict.

Nothing here is invented. Here's exactly what happens when you run a check.