GEO audit: see your site the way ChatGPT, Gemini and AI crawlers do

A GEO audit, sometimes called an AI SEO audit, is a site check for generative engine optimization. It asks whether AI systems can reach your pages, make sense of what you sell and lift a quotable answer from them.

A classic SEO audit asks whether Google can crawl and index you. A GEO audit adds what answer engines and agents need: explicit rules for AI crawlers, content that answers before it sells, structured data that says what each page is, and machine-readable entry points. LogNorm runs both. The site audit covers classic SEO; this page covers the GEO side.

It works on two levels. Up to 300 key pages are inspected live in full, and up to 500 indexable content pages are read by AI for answerability and entity clarity.

Scores of 80 and above read AI-ready, 60 to 79 Room to grow, 40 to 59 Needs work and below 40 Hard for AI to use. Two problems cap the score whatever else you fix: a noindex on your home page caps it at 60, and blocking most AI answer-engine crawlers caps it at 45.

The audit lists 14 AI crawlers and what your robots.txt tells each one, then tests the home page with their user agents to catch a CDN or firewall that blocks them anyway. Hover any status to see the exact robots.txt lines behind it.

• Answer crawlers fetch pages to cite in live answers, such as OAI-SearchBot, Claude-SearchBot and PerplexityBot. Blocking them is critical.
• On-request crawlers open a page when a user asks, such as ChatGPT-User and Perplexity-User.
• Training crawlers collect data for model training, such as GPTBot, ClaudeBot, Google-Extended and CCBot. Blocking them is only a notice, and a deliberate choice.

An AI read scores each content page from 0 to 10 on both, and a page passes at 7. Answerability asks whether an engine can lift a direct, self-contained answer; entity clarity asks whether the page makes plain what you are, your category and who it's for.

The read also flags each page Answer first or No direct answer, Quotable or Nothing to quote, and Stale when it presents old years, versions or prices as current. These findings carry a LogNorm's call label with a confidence, because they are judgements, not measurements. Their moves retire only after two audits in a row without the issue.

It checks whether /llms.txt and /llms-full.txt exist and, for llms.txt, how useful it is. An HTML page served at that address doesn't count.

LogNorm counts links and sections, rates quality out of five and lists the problems: a missing # Name title, no > summary line, no ## sections, most links pointing off-site, or links without descriptions. No usable llms.txt is a warning; no llms-full.txt is a notice. Both can be generated from your crawled pages with no AI call.

The Pages tab lists every audited page with its type, GEO score, answerability, entity clarity and issue counts. Sort by Hardest to quote first or Least clear first to find the pages worth rewriting.

Click a page to see every check with its evidence and the fix. Page checks cover content that needs JavaScript, noindex tags and headers, invalid or missing JSON-LD for the page type, missing FAQPage schema, comparison pages without a table, missing author or date, and canonicals. For the pages hardest to quote, findings include suggested rewrites marked Try. All URLs (CSV) downloads every affected page for a check.

Entity clarity is scored on every page but only fails commercial pages such as home, product, pricing, comparison and alternatives. Articles and docs show it as a meter.

The agent readiness card checks the machine-readable entry points agents look for. The checks mirror public scanners such as isitagentready.com, so fixing them here moves those results too.

• Markdown for agents: your home page returns text/markdown when asked with Accept: text/markdown.
• A Content-Signal line in robots.txt.
• An RFC 8288 Link header pointing at llms.txt, an API catalog, docs or the sitemap.
• An API catalog at /.well-known/api-catalog.
• MCP server card, agent card, Agent Skills index or AI catalog files, scored as one check.

If you have nothing for agents to call, you can skip the protocol files. Informational checks such as OAuth discovery never affect your score.

Critical and warning findings become moves on your backlog automatically, and Generate fix builds the asset that fixes a check when LogNorm can. Assets include llms.txt, llms-full.txt, a robots.txt AI section, an agent discovery kit, a JSON-LD bundle and answer-first pages.

After you ship, Validate fix re-inspects only the affected pages; site-wide checks re-run the whole audit. Your coding agent can do the whole loop: claim the move, change the files in your repo and ask LogNorm to re-check. See the AI SEO agent page. The audit re-runs weekly after the first one, so the score history fills in.