AI Visibility

How to Audit Your Website for AI Search Visibility

A repeatable audit for separating crawler access, search eligibility, AI mentions and citations, and actual website visits.

Four lenses examining a website structure, representing distinct stages of an AI visibility audit.
Key takeaway

To audit AI search visibility, check four things separately: whether relevant crawlers can access your pages, whether those pages meet the platform’s eligibility requirements, whether answers actually mention or cite you, and whether visitors reach your site. None of these signals proves the next one.

Start with a small, repeatable sample. Record what you checked, what you observed and what remains unknown before deciding whether to fix a technical issue, improve content or invest in ongoing visibility tracking.

Your 30-minute AI visibility audit

  1. Minutes 0–5: Pick the sample. Choose three to five important public pages and five questions your audience might ask. Include a mix of problem, comparison and branded questions. Record the date, platform, market and language.
  2. Minutes 5–12: Check access. Review robots.txt and any CDN or firewall rules for Googlebot and OAI-SearchBot. Note intentional blocks separately from suspected errors. For priority Google pages, use Search Console’s URL Inspection tool to investigate access and indexing. Google’s AI features guidance and OpenAI’s crawler documentation describe the relevant controls.
  3. Minutes 12–18: Check eligibility. In URL Inspection, review Google’s indexed status and selected canonical, then run a live test if you need to investigate the current page. Check snippet restrictions on pages intended to appear as supporting links. A successful live test is not a guarantee of indexing or display. Google’s URL Inspection documentation
  4. Minutes 18–25: Observe answers. Run the same questions separately in Google’s AI features and ChatGPT search. For each available answer, record whether your brand is mentioned, whether your site is linked as a source and which URL is cited. Save the answer or a screenshot where practical; mark an unavailable answer as unavailable, not as a zero.
  5. Minutes 25–30 — Review impressions and visits. Check the Generative AI performance report in Search Console and review identifiable referrals in your website analytics. Record the date range, then assign an owner and a retest date for each issue.

Know what each audit signal tells you

SignalWhat to checkWhat it does not establish
Crawl accessrobots.txt, infrastructure rules and available crawl diagnosticsIndexing, an answer citation or a human visit
EligibilityPlatform-specific requirements; for Google, indexed status and snippet eligibilityThat a page will be selected or shown
Observed mentionBrand name in a saved answerA link to your site or a visit
Observed citationA link to your site in a saved answerThat the brand was named or that anyone clicked
VisitIdentifiable referral and landing-page data in analyticsEvery mention or citation that occurred

1. Check crawler access and eligibility by platform

Google AI Overviews and AI Mode

Google says a page must be indexed and eligible to show a snippet in Google Search to appear as a supporting link in AI Overviews or AI Mode. It lists no additional technical requirements for those links. These features can show different answers and links, so record which feature you observed rather than treating them as one result. Google’s AI features guidance

  • Inspect each priority URL in Search Console. Review the indexed result, crawl details and Google-selected canonical; use the live test to investigate the current version. Google’s URL Inspection documentation
  • Check whether robots.txt, hosting or CDN rules prevent Googlebot from accessing intended public pages. Google also identifies successful page responses and indexable content among its minimum technical requirements. Google’s AI features guidance · Google’s technical requirements
  • Review noindex and snippet controls such as nosnippet, data-nosnippet and max-snippet against your publishing intentions. Keep restrictions that serve a deliberate policy; investigate unintended ones. Google’s AI features guidance
  • Check that important information is available as text and that internal links help users and crawlers discover the page. Google’s AI features guidance

Do not use robots.txt as a privacy control. Google notes that blocked URLs can still appear in search results; password protection is an option for private content. Google’s robots.txt guide

For ChatGPT search, inspect access for OAI-SearchBot separately from your Google checks. OpenAI says sites that opt out of OAI-SearchBot will not be shown in ChatGPT search answers, although they can still appear as navigational links. Allowing the bot is an access decision, not proof that a particular page will be cited. OpenAI’s crawler documentation

Do not confuse OAI-SearchBot with GPTBot, which OpenAI describes as a crawler for content that may be used in training. Their robots.txt settings are independent. If you change OAI-SearchBot rules, OpenAI says its systems can take approximately 24 hours to adjust; schedule a later retest. OpenAI’s crawler documentation

2. Record mentions and AI citations

Create a simple audit sheet with one row per question, platform and observation. Include: For more on reading source links in tracked answers, see Ansvra’s guide to AI citations.

  • Exact question, date, market, language and platform or feature.
  • Answer available: yes or no.
  • Brand mentioned: yes or no, with a short note on context.
  • Your site cited: yes or no; if yes, record the linked URL.
  • Other cited pages worth reviewing and a link or reference to the saved observation.

Use the same question set and audience context when repeating the audit. A small manual sample describes only the answers you checked; a missing citation is a reason to investigate, not proof that your content is absent or inaccessible. Google notes that search results can vary with factors including time, location and device. Search Console’s Performance report

3. Separate visibility from AI search analytics for visits

In Search Console, use the Generative AI performance report for Search to review impressions for your pages in features such as AI Overviews and AI Mode. Compare that view with overall Web search performance for the same period. Google includes AI feature impressions in the overall figures, so keep the two views separate when reporting totals.

In your website analytics, review the referral information available for the same period and identify relevant landing pages. Label the result as identifiable referrals, not all visits influenced by AI answers. A citation is not a visit, and an observed visit does not reveal every answer a person saw before arriving. See Ansvra’s guide to traffic from AI for examples of how to interpret referral reports.

Separate trays represent crawler access, eligibility, citations and visits.
Access, eligibility, citations and visits are distinct signals.

Hypothetical B2B worked example

Scenario, not a real audit: A B2B software team wants to assess a product page, a comparison page and an implementation guide. Its sample questions include “How should a team evaluate this type of software?” and a branded question about the product.

  1. Access: The SEO lead checks Googlebot and OAI-SearchBot rules for all three URLs and logs any suspected blocks for the web team to investigate.
  2. Google eligibility: The lead inspects each URL’s indexed status, selected canonical and snippet controls. Any unexpected finding becomes a separate issue with an owner and retest date.
  3. Observed answers: The content lead runs the questions in Google AI features and ChatGPT search, recording mentions and citations independently. If an answer is unavailable, the row remains unavailable.
  4. Visits: The analytics lead reviews Google Search performance and identifiable referrals to the three pages for the same period, without treating either as a complete count of AI visibility.
  5. Decision: The team prioritizes verified access or eligibility problems first, then reviews cited pages and audience questions for content opportunities. It records findings only after the checks are complete.

When to consider a visibility tracking plan

A manual audit is a useful starting point. Consider a tracking plan when your team needs to review a larger, stable set of questions across platforms, markets or languages; compare observations over time; or give multiple stakeholders access to the underlying answers.

Before choosing a plan, define your question set and reporting scope. Check how it distinguishes mentions from citations, handles unavailable answers, preserves answer evidence and reports traffic separately. Agree on a review cadence and a person responsible for acting on findings. Tracking can help organize observations, but it does not replace technical checks or establish that a citation generated a visit. The AI Search Analytics overview explains how Ansvra reports visibility across tracked questions.

Short audit checklist

  • Choose priority URLs, questions, platforms and audience context.
  • Check Googlebot and OAI-SearchBot access separately.
  • Inspect Google indexing, canonicals and snippet controls.
  • Save observed answers and distinguish mentions from citations.
  • Review Google Search performance and identifiable referrals separately.
  • Assign owners, document uncertainty and schedule a retest.

FAQs

Does allowing an AI crawler guarantee a citation?

No. Access is one check, not evidence that an answer mentioned or linked to your page. For Google, even an indexed URL is not guaranteed to appear in search results. Google’s URL Inspection documentation

What can Search Console tell me about visibility in Google’s AI features?

Search Console offers a dedicated report for impressions in generative AI features. Use it to see which pages appeared. Use your website analytics to review visits you can attribute to an AI referral.

Which crawler matters for ChatGPT search?

OpenAI identifies OAI-SearchBot as its search crawler and GPTBot as a separate training-related crawler. Review each setting against your organization’s goals; allowing GPTBot is not the same decision as allowing OAI-SearchBot. OpenAI’s crawler documentation

Sources

  1. Google Search Central — AI features and your websiteAccessed
  2. OpenAI — Overview of OpenAI crawlersAccessed
  3. Google Search Console — Performance reportAccessed
  4. Google Search Central — Introduction to robots.txtAccessed
  5. Google Search Central — Technical requirementsAccessed
  6. Google Search Console — URL Inspection toolAccessed
  7. Google Search Central — Generative AI performance reportsAccessed

Discussion

Join the discussion

Start the conversation

Share a question or tell us what worked for you.

Comments appear after review. How we handle comments

The Ansvra journal

Keep reading

Browse all articles

Ansvra

A clearer next step.

MEET YOUR AI GUIDE

A little clarity.
A clearer next move.

Ask about Ansvra, AI visibility, or your next step with content.

Opening chat…