AI Visibility Audit Checklist for Digital Marketers
Many brands rank well in Google and still go missing from ChatGPT, Gemini or Copilot answers. A classic SEO audit won’t show why, because it never looks at the answers themselves.
Below are 27 checks, each with a pass or fail condition. When we ran the citation and brand checks on our own site, most citations went to a handful of pages.
The checks cover ChatGPT, Google AI Overviews, AI Mode, Gemini, Microsoft Copilot and Claude.
Key takeaways
- An AI visibility audit checks the answers themselves: who gets named, which URLs get cited and whether your brand is described correctly.
- Start with technical access. If Googlebot or an AI crawler can’t fetch a page, that page can’t be cited.
- Google and Microsoft now report AI visibility directly, in the Search Console Generative AI report and Bing’s AI Performance report.
- Track mentions and citations separately, since AI engines can cite your pages without naming your brand.
- Citations bunch up on a few pages. On our site, one list of GitHub repositories earned 7,007 citations by itself, so map your cited URLs before writing anything new.
- Every failed check needs an owner, a target URL and a recheck date, or the audit becomes a report nobody uses.
What is an AI visibility audit?
An AI visibility audit is a structured review of how AI answer engines find, describe and cite your brand. It pairs a technical check of your site with a test of real AI answers.
A classic SEO audit asks whether your pages can rank. This one also asks whether AI answers mention you and which of your URLs they cite. When they skip you, it shows which competitors or third-party sites they trusted.
The two audits overlap. Google says its AI features are rooted in its core Search ranking and quality systems, so SEO basics still matter.
What deliverables should come from an AI search audit?
A score on its own changes nothing. Plan for six deliverables your team can work from:
A prompt set of 25 to 100 buyer prompts, with current mention and citation rates per platform as the baseline.
A map of cited sources: your URLs, competitor URLs and the third-party domains AI engines rely on.
A technical access report on crawler rules, firewall behavior, indexability and AI feature settings.
An answer accuracy review of how each platform describes your brand, product category and audience.
A page fix list naming the priority pages and the content or structure changes each one needs.
An action plan that gives each fix an owner, a target URL, a priority and a recheck date.
The AI visibility audit checklist: 27 pass/fail checks
Mark each check pass or fail. If you can’t tell, it’s a fail until you have evidence.
1. Technical access
Can search engines and AI crawlers fetch your main pages and read what’s on them?
- AI search crawlers are allowed. Pass: robots.txt lets Googlebot, Bingbot, OAI-SearchBot, Claude-SearchBot and PerplexityBot crawl the sections you care about. Fail: any of them is blocked without a decision.
- Firewall and CDN let bots through. Pass: priority URLs return a 200 status to AI user agents. Fail: bots get 403 or 429 errors, or challenge pages.
- Priority pages are indexable. Pass: no noindex tag, a correct canonical URL, and inclusion in the XML sitemap. Fail: any of the three is missing.
- Main content is in the HTML. Pass: the main text appears in the page source without running JavaScript. Fail: the source shows an empty shell.
- Google’s AI feature setting is deliberate. Pass: the Search generative AI control in Search Console matches your decision. Fail: nobody has checked it.
- Bing has indexed your priority pages. Pass: URL inspection in Bing Webmaster Tools shows them as indexed. Fail: some are missing, which limits Copilot.
Google added the Search generative AI control for all sites on August 31, 2026. It lets you include or exclude your site from AI Overviews and AI Mode. Inclusion is the default.
Our step-by-step guide shows how to check if AI crawlers can access your website, with robots.txt examples for every major bot.
2. Content and answer readiness
Here the question is whether a page gives an AI engine something specific enough to quote.
Each priority prompt has one target URL. Pass: every tracked prompt maps to a single page. Fail: several pages compete for the same question, or none answers it.
The answer comes first. Pass: the first two sentences of the matching section answer the question directly. Fail: the answer is buried after background text.
Headings match real questions. Pass: H2s use the wording from Search Console queries and tracked prompts. Fail: headings are clever but vague.
Claims are specific and sourced. Pass: the main claims rest on a number with a date or a named source. Fail: the page relies on general statements.
The page adds something new. Pass: it brings first-hand data or a point of view of its own. Fail: it restates what every competitor already says.
Facts are current. Pass: the page shows an updated date and facts were reviewed in the last 12 months. Fail: prices or statistics are out of date.
Google’s guide to optimizing for generative AI search calls this non-commodity content. It also says Google Search doesn’t use llms.txt files, so a missing llms.txt isn’t a fail for Google.
3. Entity and brand signals
Does each AI engine know who you are, what you sell and who it’s for? Well-defined brand entities raise the odds that an answer names you as well as citing you.
- One consistent brand description. Pass: your homepage, About page and main profiles describe your category and audience the same way. Fail: the descriptions conflict.
- Company facts are visible. Pass: About, contact and leadership details are on the site, and Organization markup matches them. Fail: the facts exist only in markup, or nowhere.
- AI describes you correctly. Pass: asked "What is [your brand]?", each platform names your category, audience and main use. Fail: wrong category or outdated facts.
- You’re named where you’re cited. Pass: answers that cite your pages also mention your brand. Fail: your URLs support an answer that recommends competitors.
- Comparison pages exist. Pass: you publish fair comparison and alternatives pages for the competitors buyers name. Fail: AI engines rely only on other sites’ comparisons.
4. Citations and third-party sources
Which pages do AI engines already trust for your topics? Count your own and everyone else’s.
- You know your cited URLs. Pass: you have a list of your cited URLs by platform and prompt. Fail: you only know that "we show up sometimes."
- Citations reach commercial pages. Pass: pricing and comparison pages earn citations for buying prompts. Fail: one blog post carries almost all citations.
- Competitor sources are mapped. Pass: you know which competitor URLs win your priority prompts and what they offer. Fail: competitors are tracked by name only.
- You appear on trusted third-party sites. Pass: you have a genuine presence on the review sites and communities AI engines cite. Fail: they never mention you.
- Lost citations are reviewed. Pass: each month someone checks the URLs that lost citations and why. Fail: losses go unnoticed until traffic drops.
Google warns that chasing inauthentic mentions doesn’t help. Earn that presence with real customers and real data. Citation monitoring shows the cited domains and exact URLs for each tracked prompt.
5. Measurement
Fix something without these in place and you won’t be able to prove it worked.
- A fixed prompt set. Pass: 25 to 100 prompts across awareness, comparison and purchase stages, rerun on a schedule. Fail: ad hoc checks in a personal browser.
- A baseline per platform. Pass: mention rate, citation rate and share of voice are recorded for each platform. Fail: one blended score hides platform gaps.
- Search Console AI data is reviewed. Pass: someone reviews AI Overviews and AI Mode impressions by page each month. Fail: the report has never been opened.
- Bing AI data is reviewed. Pass: you check Copilot citations and grounding queries in Bing Webmaster Tools. Fail: Copilot isn’t measured.
- AI traffic and conversions are tracked. Pass: analytics separates visits from AI platforms and links them to sign-ups or sales. Fail: AI visits sit inside referral traffic.
Search Console’s Generative AI performance report shows impressions in AI Overviews and AI Mode by page, country and device. It shows no queries or clicks, and Google rolled it out to all sites on August 31, 2026.
Microsoft’s AI Performance report in Bing Webmaster Tools shows citations, cited pages and grounding queries. It covers Copilot and Bing AI summaries.
Neither report covers ChatGPT, Gemini or Claude, so you still need your own prompt tracking. Our guide to AI visibility KPIs and benchmarks explains which numbers to report and how to calculate AI share of voice.
How to score your AI visibility audit checklist
Count the passes in each area: technical access out of 6, content out of 6, and entity, citations and measurement out of 5 each.
| Area | Checks | Fix order |
|---|---|---|
| Technical access | 6 | First |
| Measurement | 5 | Second |
| Content and answer readiness | 6 | Third, by commercial value |
| Entity and brand signals | 5 | Third, by commercial value |
| Citations and third-party sources | 5 | Third, by commercial value |
- Fix technical fails first. A blocked or unindexed page can’t be cited, however good its content.
- Set up measurement next. Without a baseline, you can’t tell whether later fixes worked.
- Then sort the remaining fails by commercial value. A fail on a pricing or comparison prompt matters more than one on a definition prompt.
Recheck the same prompts two to four weeks after each fix. AI answers vary from run to run, so look at the trend across several runs. A single answer proves little.
A worked example: what our own audit found
We ran the citation and brand checks on our own site. Ansvisor, an open-source AI search intelligence platform, already tracks prompts about AI search, so the data was there.
The figures below are Ansvisor tracking data, September 15 to October 14, 2026.
- One page did most of the work. A list of GitHub repositories for AEO and GEO earned 7,007 citations on its own.
- Citations and mentions didn’t move together. Brands with similar citation counts were named at very different rates.
- Third-party sites carried real weight. Among third-party domains, reddit.com held a 14.5% citation share, youtube.com 11.6%, github.com 11.5% and linkedin.com 10.4%.
So our own fix list starts with entity work. That means clearer brand statements on cited pages and more genuine presence on the communities AI engines cite.
How to run a first AI visibility audit in five days
Never measured AI visibility before? This plan gets you a baseline in one working week.
- Day 1: build a prompt set from Search Console queries, sales questions and competitor comparisons.
- Day 2: run every prompt on each platform and log mentions, cited URLs and competitors.
- Day 3: complete the technical access checks for your top 20 pages.
- Day 4: review the content and entity checks on the pages that should win your priority prompts.
- Day 5: score the checklist and turn each fail into a task with an owner and a recheck date.
For Day 1, our guide to prompt research for AEO and GEO shows how to find the questions worth tracking. If you sell audits to clients, read how agencies use AEO audits to win more clients.
How Ansvisor helps you run the audit
Ansvisor runs your prompt set on a schedule across ChatGPT, Google AI Overviews, AI Mode, Gemini, Copilot and Perplexity, and keeps every answer. Claude tracking is available on Enterprise plans.
The AI Visibility Site Audit scores any public URL against 47 weighted AEO and GEO signals, which covers most technical and content checks above. Competitor benchmarking shows the prompts where rivals get named and you don’t.
FAQ
What deliverables should come from an AI search audit?
Expect a tracked prompt set with a visibility baseline, a citation and source map, and a technical access report. A full audit also includes an answer accuracy review, a page fix list and an action plan with owners and recheck dates.
What does an initial AI visibility audit look like?
It starts with 25 to 100 prompts that buyers ask, run on ChatGPT, Google AI Overviews, AI Mode, Gemini, Copilot and Claude. You log mentions, cited URLs and competitors, then check crawler access on your main pages. You end up with a baseline to compare against later.
How is an AI visibility audit different from an SEO audit?
An SEO audit checks whether pages can be crawled and indexed, and how they rank and earn links in classic results. An AI visibility audit adds the answers themselves: brand mentions, cited URLs, winning sources and whether AI describes you correctly.
What tools do you need to offer AI visibility audits?
You need Google Search Console and Bing Webmaster Tools for access and AI reports, a site audit tool, and analytics for AI referral traffic. For repeatable prompt tracking across platforms, add an AI visibility tracker that stores answers and citations over time.
How often should you run an AI visibility audit?
Run a full audit once a quarter and after any major site change, product launch or repositioning. In between, track your priority prompts weekly, because AI answers and cited sources change more often than organic rankings.
Cihan Geyik
Co-founder at Ansvisor
Cihan Geyik is the co-founder of Ansvisor, an open-source, cloud-ready AI Visibility platform for AI Search. With more than 15 years of experience in digital marketing and growth, he writes about AI visibility, AI search, AEO, GEO, citations, and answer engines. He focuses on helping brands understand and improve their presence across ChatGPT, Gemini, Perplexity, Google AI Overviews, and other AI-powered discovery platforms.
Let AI Search work for you
Each prompt you win brings more mentions, more citations and more buyers. Ansvisor keeps finding the next one.
Get more Ansvisor insights in Google
Add Ansvisor as a Preferred Source to see more of our AI Search, AEO, GEO, and AI Visibility insights across Google.
Make Your Brand Visible in AI Answers.
INDUSTRIES
© 2026 Ansvisor. All rights reserved. Ansvisor is an open-source AI Search Intelligence Platform for AI Visibility. Ansvisor is a brand owned and operated by Empler AI Inc.