AI visibility is the degree to which your site can be discovered, understood, and cited by AI answer engines. This checklist covers the twelve items worth auditing, in the order that makes practical sense. Work top to bottom: technical gates first, then structure, then authority, then monitoring.
None of these items guarantees a citation. They remove common reasons a site is skipped and make your content easier to use as a source. Run the free check at https://ai.ynnadmusic.uk at the start to catch crawler-level problems before you invest elsewhere.
The 12-point checklist
- 1. Audit your robots.txt for AI crawlers. Look for
Disallowrules attached to AI user-agents such asGPTBot,OAI-SearchBot,PerplexityBot,ClaudeBot, andGoogle-Extended. Decide deliberately for each. Remember that an unmentioned crawler is allowed by default. Blocking search agents likeOAI-SearchBotandPerplexityBotis the most likely to reduce discoverability. See the AI crawler list and the GPTBot guide.
- 2. Set Google-Extended explicitly. Understand that
Google-Extendedcontrols specific AI training uses, not Googlebot or AI Overviews. Choose allow or block according to your policy, and do not expect it to change your search visibility. Read Google-Extended explained.
- 3. Confirm pages are crawlable and indexable. Key pages should return HTTP 200, sit behind no login or paywall, and carry no
noindexdirective. Check that general crawling is not blocked. This is the foundation for both search and answer engines.
- 4. Publish and submit a sitemap. Your
sitemap.xmlshould list canonical URLs and be submitted to the major search consoles. Keep it generated from a single source of truth so it does not drift from your actual pages.
- 5. Add an llms.txt. Place a short, machine-readable summary at your site root listing your most important pages. It is not a ranking factor and its support is inconsistent, but it is free and takes ten minutes. See the llms.txt guide for a copy-paste example. It does not override robots.txt.
- 6. Add structured data. Use JSON-LD with schema.org types such as
Article,FAQPage,Organization, andBreadcrumbList. Structured data supplies explicit facts a machine can extract. Validate it and keep it accurate; it improves clarity but does not force citation.
- 7. Answer the query in the first paragraph. Each page should open with a direct, self-contained answer to its main question. An engine scanning for a response should find one immediately, without reading the entire page.
- 8. Use descriptive headings and structure. Headings that phrase real questions give clear extraction anchors. Break content into short paragraphs, bullets, and tables. Keep one idea per section and make important text part of the HTML, not JavaScript-only.
- 9. Clarify your entities. State consistently who you are, what each page covers, and how concepts relate. Include clear author and organisation information. Consistent naming across your site and profiles helps a system build a reliable picture.
- 10. Add an FAQ and keep content fresh. Genuine question-and-answer pairs are easy to extract; mark them up with
FAQPageschema where appropriate. Show accurate publish and update dates, and review content regularly — freshness matters for many queries.
- 11. Earn third-party mentions and internal links. Reputable mentions and links from other sites are among the strongest trust signals, and they take time to earn. Internally, link related pages into clear topic clusters so crawlers can discover and understand your coverage. Avoid paid link schemes.
- 12. Monitor and re-audit. There is no complete analytics console for AI citations. Search your key questions in ChatGPT, Perplexity, and Google periodically, note whether your domain appears, watch referrer traffic, and re-run the checker after any migration, plugin update, or CDN change. Revisit the checklist every quarter.
How to use this audit
Score each item as done, partial, or missing. Fix the missing technical items first — a blocked crawler makes downstream work pointless. Then address structure and content, which usually offer the biggest gains for the least cost. Finally, treat authority and monitoring as ongoing work rather than one-time tasks.
If you want a starting point for the content side, the guides on getting cited in AI Overviews and getting cited by Perplexity go deeper on individual tactics.
Frequently asked questions
Does completing this checklist guarantee AI citations?
No. It improves the conditions under which a citation is possible, but selection depends on the specific query, available sources, and signals that are not fully public. Think of it as raising your odds, not flipping a switch.
Which items matter most?
Crawler access and indexability come first — if a system cannot reach or index your pages, nothing else helps. After that, answer-first content and clear structure usually give the biggest practical return.
How often should I re-run the audit?
At least quarterly, and immediately after a site migration, redesign, plugin update, or CDN change. These events commonly regenerate robots.txt or introduce noindex tags.
Is this the same as a technical SEO audit?
It overlaps heavily, but AI visibility places more weight on extractability — direct answers, clear structure, and entity clarity — because answer engines lift specific passages rather than just ranking whole pages.