AI Search Visibility Audit Checklist

An AI search visibility audit answers a blunt question: when a buyer asks an AI system about your category, does your brand show up with evidence?

Traditional SEO checks still matter. Google still needs crawlable pages, useful content, accurate canonical tags, and clear structured data. The Google SEO Starter Guide is still a good baseline. But AI answer engines add a second layer. They summarize, compare, and cite. That means a page can be indexed and still fail the answer test.

Use this checklist before you rewrite pages. It separates technical access problems from content gaps, citation gaps, and competitor gaps.

1. Confirm The Pages Can Be Found

Start with the boring checks. If your public pages are blocked, missing from the sitemap, or canonicalized to the wrong URL, the rest of the audit becomes guesswork.

Check:

• Important public pages return HTTP 200.
• Canonical tags point to the same URL users and crawlers should index.
• The XML sitemap includes homepage, use-case pages, comparison pages, and core blog guides.
• Robots rules do not block public pages.
• Private routes such as dashboard, onboarding, report shares, and auth pages stay blocked where appropriate.

For AEO Table, the public layer includes the homepage, the AI search monitoring use case, the competitor AI visibility use case, and core guides such as the AI search visibility baseline.

2. Check AI Crawler Access

AI search visibility depends partly on which systems can fetch and understand your public pages. Do not treat every crawler as the same thing.

OpenAI documents separate crawler roles, including OAI-SearchBot and GPTBot, and says webmasters can manage them with robots.txt rules in its crawler documentation. Perplexity also documents PerplexityBot and recommends allowing it if you want Perplexity to surface and link your pages in search results, according to its crawler docs.

Audit:

• Does /robots.txt allow the AI search crawlers you care about?
• Are you accidentally blocking public content through CDN or WAF rules?
• Does the site have a useful /llms.txt file that points AI systems to important public resources?
• Are canonical URLs and sitemap URLs consistent with the URLs listed in /llms.txt?

Robots rules are not a ranking strategy. They are an access policy. Keep them explicit, documented, and aligned with your risk tolerance.

3. Build A Stable Query Set

Do not audit with random prompts. Build a small query set that mirrors buyer research.

Use five groups:

1. Category questions: "What are the best tools for tracking AI search visibility?"
2. Problem questions: "How do I know if ChatGPT mentions my brand?"
3. Comparison questions: "AEO Table vs manual AI visibility tracking."
4. Proof questions: "Which sources do AI answers cite for this category?"
5. Risk questions: "Why does my brand not appear in AI answers?"

If you need a deeper query framework, use the AI search query set guide. Keep the core questions stable so your next audit can be compared against this one.

4. Record Mentions, Not Just Presence

Presence is a weak metric by itself. A brand mention can be strong, weak, neutral, or harmful.

For each answer, record:

• Brand mentioned: yes or no.
• Mention position: early, middle, late, or passing reference.
• Framing: recommended, listed, caveated, ignored, or compared.
• Competitor mentions: which alternatives appear and how they are framed.
• Source support: whether the answer cites your domain, a third-party source, or no source.

This is where AI visibility starts to differ from rank tracking. The answer text matters as much as the citation.

5. Audit Citations And Source Quality

Do not stop at "we were cited." Review the source.

Classify every cited URL:

• Owned source: your website, docs, blog, pricing, comparison, or help page.
• Earned source: analyst, media, review site, directory, partner page, or community thread.
• Competitor source: competitor pages or competitor-favorable comparisons.
• Weak source: outdated, thin, off-topic, or inaccessible.

Google says structured data can help it understand page content, but it does not guarantee rich results. The same practical rule applies to AI visibility: schema helps, but the source page still needs useful content. Use Google's structured data introduction and FAQ structured data guidance as guardrails, not as a shortcut.

6. Find Competitor Gaps

Competitor visibility is often the most useful part of the audit.

Look for:

• Competitors that appear when your brand is absent.
• Competitors cited by stronger third-party sources.
• Competitors framed as safer, cheaper, more complete, or more established.
• Queries where competitors appear in multiple channels but you appear in none.

For a repeatable workflow, use the competitor AI visibility use case. It keeps competitor names, aliases, and domains stable so each Run is comparable.

7. Turn The Audit Into A Backlog

The audit should end with actions, not a dashboard.

Create four buckets:

• Technical access fixes: robots, sitemap, canonical, blocked assets, missing /llms.txt.
• Page refreshes: update pages that should answer high-intent questions.
• New content: publish comparison, proof, template, and FAQ-led pages where buyers need direct answers.
• Third-party proof: earn or update sources that AI answers already trust.

Prioritize by buyer impact. A missing brand mention on a high-intent comparison query matters more than a weak citation on a broad awareness prompt.

Audit Checklist

Use this as the short version:

• Public pages return 200 and self-canonicalize.
• Sitemap contains the public pages that should rank or be cited.
• Robots allows public content and blocks private surfaces.
• AI search crawlers are handled intentionally.
• /llms.txt points to the most important public pages.
• Query set covers category, comparison, proof, and risk questions.
• Answers are captured with full text, not just scores.
• Brand mentions and competitor mentions are classified.
• Citations are grouped by source type and quality.
• Findings map to a content, technical, or proof backlog.

The Bottom Line

An AI search visibility audit is not a one-time prompt test. It is a controlled inspection of access, answers, citations, competitors, and next actions.

Start with a narrow audit. Then turn the best questions into a repeatable AI search visibility baseline. Once the baseline exists, every new Run tells you whether the work improved visibility or only created more pages.

Create a free AEO Table account to run your first AI visibility baseline with repeatable Tasks and shareable evidence.