GEO audit checklist: 26 checks for AI search visibility

What a GEO audit covers, and what it cannot

A GEO audit checks the technical path between your page and an AI answer engine. The engine's crawler has to be allowed in, the page has to be indexable, and the text has to be readable without guesswork. Google states that there are no extra technical requirements for its AI features beyond ordinary Search: a page must be indexed and eligible to show a snippet. The same basics matter when ChatGPT, Claude or Perplexity search the web: their search crawlers fetch pages, and a blocked or unreadable page gives them nothing to cite.

An audit cannot tell you whether an engine will mention your brand. Relevance, competition and the question a person asks decide that. Treat the audit as the part you control, and measure mentions separately (see how to track brand mentions in AI search).

The checklist below has 26 checks in five groups. Nineteen are weighted and add up to 100 points. Seven are reported for information with zero weight, because they reflect a site owner's choice rather than a readiness problem.

Crawler access: 11 robots.txt checks

Each crawler is checked against your published robots.txt for the audited path. Search crawlers carry points because blocking them removes you from that engine's answers. Training crawlers carry none: allowing or blocking them is a business decision.

Fix: a bot with its own User-agent group ignores the * group, so check named groups first. Our AI crawler list explains each token, and the robots.txt generator builds a fragment from your choices. Remember that robots.txt is only one layer: a CDN or firewall rule can still block a crawler that robots.txt allows.

Access and indexing: 6 checks, 38 points

Fix: a stray noindex is the most expensive failure on this list, because it removes the page from the index that AI features draw on. Look for it in your SEO plugin, staging settings that leaked to production, and server headers. A single-language site can ignore the hreflang check; the audit reports it so multilingual sites notice missing alternates.

Structured data and content: 7 checks, 36 points

Fix: start the page with a short paragraph that answers the page's main question directly, then support it. Keep structured data consistent with the visible text; Google's guidelines say not to mark up content readers cannot see. Our guide to schema markup for AI search covers which types are worth adding. The 200-word threshold is an inspection prompt, not a ranking rule: do not pad a page to reach it.

Rendering and optional signals: 2 checks

• Content in initial HTML (6 points). The audit reads the raw HTML without running JavaScript. If the page is a near-empty app shell, crawlers that do not render JavaScript see nothing to cite. Render the main heading and text on the server.
• Optional llms.txt (0 points). Reported only. llms.txt is an optional navigation proposal, not a ranking or access control. Read does llms.txt work? before spending time on it.

A copyable GEO audit template

Paste this into a ticket or spreadsheet and work through one page at a time:

GEO audit — page: ____________  date: __________

Crawler access (robots.txt and CDN)
[ ] OAI-SearchBot allowed        [ ] Claude-SearchBot allowed
[ ] PerplexityBot allowed        [ ] Googlebot allowed
[ ] Bingbot allowed              [ ] CDN / firewall checked
[ ] Training bots (GPTBot, ClaudeBot, Google-Extended,
    CCBot, Bytespider, Applebot-Extended): decision recorded

Access and indexing
[ ] HTTP 200 on the canonical URL
[ ] No noindex in X-Robots-Tag    [ ] No noindex in meta robots
[ ] Canonical URL present         [ ] hreflang (multilingual only)
[ ] XML sitemap readable and listed in robots.txt

Structured data and content
[ ] JSON-LD parses and matches visible facts
[ ] One H1, no skipped heading levels
[ ] Early answer paragraph (25–120 words)
[ ] Visible FAQ where it helps readers
[ ] Author and date where the page type needs them
[ ] Main text present in the initial HTML

Recheck date after fixes: __________

How to run a GEO audit

1. Open https://your-site.com/robots.txt and read every group that names a crawler from the table above.
2. Run curl -sI https://your-site.com/page to see the status code and any X-Robots-Tag header.
3. View the page source, not the rendered page, and search for noindex, canonical, hreflang and application/ld+json.
4. Confirm the first paragraphs and headings are in that source.
5. Fix the failures in order of points, then recheck the same page.

Or run the free AI visibility checker on the page: it performs these 26 checks on the public HTML and returns each failure with a copyable fix after an email unlock.

Sources

• Google Search Central: AI features and your website
• Google Search Central: Robots meta tag and X-Robots-Tag specifications
• Google Search Central: How to specify a canonical URL
• Google Search Central: Localized versions of your pages (hreflang)
• Google Search Central: General structured data guidelines
• Sitemaps.org: Sitemap protocol
• RFC 9309: Robots Exclusion Protocol
• OpenAI: Overview of OpenAI crawlers

Reviewed on 2026-10-09. Point values match the CiteVisor audit at that date. Passing every check means the technical path is open; it does not guarantee that any engine will mention or cite a page.