What the audit checks
Eleven checks, weighted by how much each one decides whether AI can use the page. Notes on training crawlers and Cloudflare come with the results but don't change the score.
| Check | Points | Why it matters |
|---|---|---|
| The page loads for crawlers | 20 | A page that answers with an error, a redirect loop or a bot challenge gives crawlers nothing to read. |
| AI search and answer crawlers may read it | 15 | ChatGPT-User, OAI-SearchBot, Claude-User, Claude-SearchBot, PerplexityBot and Perplexity-User fetch pages for AI answers and AI search. |
| The main text is in the HTML | 15 | Text that only appears after JavaScript runs may be invisible to AI crawlers. |
| Google and Bing may read it | 10 | Gemini, Google AI Overviews and AI Mode use Google's index, and Microsoft Copilot uses Bing's. |
| It has structured data | 10 | Schema.org JSON-LD states who you are and what the page holds, in a form machines read without guessing. |
| Search engines may index it and quote it | 5 | noindex keeps a page out of Google, and nosnippet stops Google quoting it in AI Overviews. |
| It has a title and a description | 5 | They say in a line what the page is about. |
| It has one main heading | 5 | One H1 that names the topic makes the subject of the page clear. |
| It names its main address and language | 5 | A canonical URL and a lang attribute stop copies competing and tell crawlers the language. |
| robots.txt points to a sitemap | 5 | A Sitemap line in robots.txt helps every crawler find your other pages. |
| It has an llms.txt | 5 | A short markdown map of your best pages for AI assistants and agents. The least important check: Google says it doesn't use it. |
How to read the score
- 85 to 100: ready for AI search. Crawlers can reach the page, read it and tell what it is.
- 60 to 84: a few fixes away. AI can use the page, but something holds it back, often missing markup or thin HTML.
- Below 60: hard for AI to use. Usually a blocked crawler, a page built only by JavaScript, or a page that doesn't load for bots.
After the audit
Fix the first item on the list, then run the audit again. For robots.txt problems, the robots.txt checker shows the line that decides each crawler and lets you test a change, and the robots.txt generator writes a new file. The llms.txt generator drafts an llms.txt from your sitemap. The AI crawler guides explain what each crawler does.
The audit shows whether AI can read a page. To see whether assistants actually mention you, compare AI visibility tools, or see how OneLence AI visibility works: with the SEO & GEO Pack it asks your customers' questions on ChatGPT, Gemini, Perplexity, Claude, Google AI Overviews and AI Mode, shows which pages AI crawlers read, and which visitors AI assistants send you and whether they buy.
Frequently asked questions
What is an AI visibility audit?
A check of whether AI assistants and AI search can find, read and understand your pages: whether their crawlers may fetch a page, whether its text is in the HTML, whether it is marked up and whether search engines may index it. This audit scores one page out of 100 and lists the fixes, most important first.
Is this a GEO audit or an AEO audit?
Both terms mean much the same thing. GEO, generative engine optimization, and AEO, answer engine optimization, are about appearing in AI answers. This audit covers the technical part every GEO or AEO plan starts with: whether AI can reach and read your pages at all.
Does a high score mean ChatGPT will mention my brand?
No. A high score means AI crawlers can reach, read and understand the page. Whether assistants mention you also depends on what your pages say, what other sites say about you and what people ask. To see the answers themselves, you need to ask the questions on each assistant, which OneLence AI visibility does every week with the SEO & GEO Pack.
How is this different from the robots.txt checker?
The robots.txt checker goes deep on one thing: which of 13 crawlers may read a page and which line of robots.txt decides it, with an editor to test changes. The audit uses the same fetch and adds a score across access, readable text, structured data and sitemaps.
Which page should I audit?
Start with the homepage, then the pages you most want AI to quote: pricing, product pages, guides and comparisons. Each page can score differently, because robots.txt rules, rendering and markup often change from one section of a site to another.
What does the audit fetch?
The page you enter, then robots.txt, llms.txt and llms-full.txt from the site it ends up on, once each, the way a crawler would. Repeat runs within five minutes reuse the same fetch.