AI Search Visibility Audit: What Should the Report Include?

By Istiak Ahmed Soikot, Founder and CEO, Rank OG. Last reviewed: November 4, 2026.

An AI search visibility audit shows whether AI search engines can read and cite your website. The report must name the prompts tested, show dated answers, list your pages’ citations, verify crawler access, and quote Search Console’s AI data alongside what it cannot prove.

Who this is for: This guide is for business owners and marketing leads weighing an AI search visibility audit. For the implementation side, see Rank OG’s AI search optimization service.

Why is an AI search visibility audit different from a regular SEO audit?

A regular SEO audit measures things Google documents, like indexation and rankings. An AI search visibility audit has no official scoreboard, so the report must build its own: the prompts it ran, the answers it captured, and the citations your pages earned.

Google’s own guide explains the machinery: its generative AI features are rooted in core Search ranking and quality systems, using retrieval-augmented generation to pull relevant pages from the Search index and ground answers in their content with clickable links. Google also describes query fan-out, where the model generates concurrent related queries to fetch more results behind one question.

No engine publishes a ranking formula for AI citations. ChatGPT searches its own index via OAI-SearchBot, Perplexity uses its own index via PerplexityBot, and Claude searches with Claude-SearchBot, and each rewrites your question before searching. The same prompt can return a different answer tomorrow, so the report must record test conditions like a lab experiment, making results comparable.

What must the prompt set section of the report include?

The report must list every prompt word for word, the AI surface it ran on, and the date it ran. Without the exact prompts, a citation count means nothing, because a different wording can return a completely different answer.

A serious prompt set has 15 to 30 prompts written like real buyers write: “best local SEO company for dental practices” beats “SEO services.” Mix informational prompts (“how does technical SEO pricing work”) with commercial ones (“hire a technical SEO consultant”).

Test the same prompts in Google’s AI Overviews and AI Mode, in ChatGPT search, in Perplexity, and in Claude, since answers drift. No AI search visibility audit is complete on one engine alone.

How should the report document citation tracking?

Each captured answer should be saved with the full prompt, the AI surface, the date, and your cited URL. Anything less is a claim, not evidence. In an AI search visibility audit, treat citation tracking like lab work: same conditions, written down, repeatable.

Separate three things people lump together as “citations.” A citation means your URL appears as a named source. A mention means the engine names your brand but links nothing. A paraphrase means the answer uses your content without naming you. Only the first is countable, but all three show your content is in the retrieval pool.

Track your pages, not just your domain: citations spread across service pages prove your depth gets retrieved. Screenshot everything. Answers change, and a screenshot with the prompt, the surface, and the timestamp is the only proof that stands up when the numbers are questioned later.

What does Search Console’s Generative AI performance report tell you?

It is the only official measurement an AI search visibility audit can lean on. The Generative AI performance report in Search Console shows how your content performs in generative AI features on Google Search and Discover, based on real user exposure.

Manual tests show what answers say; Search Console shows what real people saw. Citations in tests but no AI impressions in Search Console usually means your prompts do not match what real searchers ask. Impressions without clicks often mean the page answers the question too completely, which is a page-level content fix.

Google’s guide states the eligibility gates: a page must be indexed and eligible to be shown in Google Search with a snippet, and the site must be included in Search generative AI features in Search Console. A page failing either gate cannot appear in AI Overviews or AI Mode, so the report should check these first.

Why does the report need a crawler access section?

AI engines can only cite pages their crawlers can read. If your robots.txt or firewall blocks their search bots, no content work will get you cited. The audit must check crawler access first, before it grades a single page.

OpenAI’s crawler documentation says OAI-SearchBot “is used to surface websites in search results in ChatGPT’s search features,” and that sites opted out of it “will not be shown in ChatGPT search answers, though can still appear as navigational links.” The settings are independent: GPTBot feeds model training, while OAI-SearchBot handles search, so a site can block training and stay citable. OpenAI also recommends allowing its published IP ranges, noting robots.txt changes take about 24 hours to reach its systems.

Perplexity’s crawler docs say PerplexityBot “is designed to surface and link websites in search results on Perplexity” Sites that want to appear should allow PerplexityBot in robots.txt and permit requests from its published IP ranges. Perplexity-User generally ignores robots.txt. The same docs include WAF configuration guidance, because a firewall can block a crawler even when robots.txt says yes.

Anthropic runs three bots: ClaudeBot collects content that may contribute to model training, Claude-SearchBot improves search result quality, and Claude-User fetches pages when a person asks. Anthropic says disabling Claude-SearchBot “may reduce your site’s visibility and accuracy in user search results,” and warns that blocking by IP address “may not work correctly or persistently.” Blanket-blocking “AI bots” to stop training is the most common accidental self-block.

Rank OG’s own site shows how easy the gap is to miss. Its robots.txt, checked November 2026, allows GPTBot, ChatGPT-User, ClaudeBot, anthropic-ai, PerplexityBot, and Google-Extended, but does not name OAI-SearchBot or Claude-SearchBot explicitly. Exact token names matter: a rule naming GPTBot does nothing for OAI-SearchBot. An audit that catches gaps like this is doing its job.

What should the report say about your content and technical setup?

For the content section of an AI search visibility audit, Google’s own guide names one lever above all others: unique, non-commodity content built on real experience. The report should grade your pages against that bar, then confirm the technical gates: indexed, snippet-eligible, and crawlable.

Non-commodity content means content only you could write. Google contrasts a first-hand review with a summary that restates what is already online. The audit should ask, page by page: does this page hold a unique point of view, first-hand evidence, or original data, or could any competitor have produced it? Pages that fail that test rarely get cited. Clear headings let engines lift the right passage.

On the technical side, the audit should verify crawlability (JavaScript content not blocked), good page experience, reduced duplicate content, and reasonably semantic HTML, which Google calls “generally a good idea” without requiring perfection.

Legitimate audit checks vs common fluff

What a real report shows	What a fluff report sells
Exact prompt list with surfaces and dates	An “AI visibility score” with no visible prompts
Screenshot evidence for every citation	Citation counts with no proof
Search Console’s Generative AI report data	No mention of Search Console at all
robots.txt and firewall audit for each AI search bot by name	Installing llms.txt, which Google says Search does not use
Non-commodity content assessment page by page	Selling content “chunking,” which Google says has no requirement
Honest limits, stated in writing	Guaranteed ChatGPT or Perplexity citations

Verdict: a report built on the left column is worth paying for, even if the numbers are modest. Run the table past any vendor before you sign.

What should an AI search visibility audit never promise?

It should never promise citations, rankings, or brand mentions in AI answers. No vendor controls another company’s retrieval step, and Google’s documentation warns against several popular “AI visibility” tactics outright.

Google’s mythbusting section is the fastest way to spot a dishonest vendor: you do not need llms.txt or other special markup, because Google Search does not use them; you do not need to break content into tiny pieces; you do not need to rewrite content just for AI systems; seeking inauthentic mentions “isn’t as helpful as it might seem”; and structured data is not required, with no special schema to add. Billing these as deliverables means charging for work Google says has no effect.

Picture a 40-page services site that orders an audit. The report finds Claude-SearchBot blocked by an old robots.txt rule, three commercial prompts driving most citations, and a pricing page that answers questions completely but earns no AI clicks. Three findings, each with evidence and a fix: that is a useful report. A fluff report would show one “visibility score” of 62 and tell you to rewrite everything for AI.

What to do next

1. Check your robots.txt for the exact search-bot tokens (OAI-SearchBot, PerplexityBot, Claude-SearchBot) and your Search Console AI settings.
2. Verify Search Console and open the Generative AI performance report to see your current real-user exposure.
3. Write down 20 prompts your buyers would actually type before you pay anyone, so you can judge any vendor’s prompt set against yours.
4. Run one round of answers yourself and save screenshots, so you have a baseline the next report must beat.
5. If the gaps look real, get Rank OG’s free AI search visibility audit and compare its report against the checklist above.

Frequently asked questions

How often should I run an AI search visibility audit?

Run an AI search visibility audit once a quarter if AI search drives a meaningful share of your discovery, and re-check the crawler section after site changes. Between audits, watch Search Console’s Generative AI performance report monthly, since it shows exposure without manual prompt testing.

Can an audit guarantee my brand will appear in ChatGPT answers?

No. No vendor controls OpenAI’s retrieval step, and OpenAI publishes no ranking formula. An honest audit measures where you stand today and what blocks your pages, then fixes those blocks. Any report that guarantees citations is selling something it cannot deliver.

Does blocking GPTBot hurt my visibility in ChatGPT search?

No. OpenAI documents GPTBot and OAI-SearchBot as independent settings: GPTBot feeds model training, while OAI-SearchBot surfaces sites in ChatGPT search answers. Blocking GPTBot only opts you out of model training. Blocking OAI-SearchBot is what removes you from search answers, so check your robots.txt for the exact token names.

How do I check whether AI search crawlers can read my site?

Open your robots.txt and confirm the exact search-bot tokens are allowed: OAI-SearchBot for ChatGPT, PerplexityBot for Perplexity, Claude-SearchBot for Claude. Then check your server logs for those tokens and verify their source IPs against each vendor’s published ranges, since user-agent strings are easy to fake.

Is an AI search visibility audit worth it for a local business?

Yes, when local queries trigger AI answers in your market. Google says generative AI responses can include local business information from Google Business Profiles. A local audit should test neighborhood-level prompts, check your profile’s completeness, and verify the same crawler gates as any other site.

Sources

• Google’s guide to optimizing for generative AI features on Google Search, Google Search Central
• Overview of OpenAI crawlers, OpenAI
• Perplexity crawlers, Perplexity
• Does Anthropic crawl data from the web, and how can site owners block the crawler?, Anthropic Privacy Center
• rank-og.com robots.txt, checked November 2026

Sources checked November 2026.

About the author: Istiak Ahmed Soikot is the Founder and CEO of Rank OG, an AI-first SEO agency working hands-on with technical SEO, local SEO, and AI search optimization.

Get a free AI search visibility audit from Rank OG and compare it against the checklist above. Start with the free audit, or reach out through the contact page.

Want the bigger picture? Read Rank OG’s generative engine optimization service page to see how audits turn into an ongoing AI search program.