json
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "What does an AI visibility audit check?",
"acceptedAnswer": {
"@type": "Answer",
"text": "An AI visibility audit checks 15 signals across three areas: technical markup (robots.txt, llms.txt, schema types), content structure (H2 format, answer placement, statistics linking), and brand authority (named author, date markup, external citations, third-party mentions). BrandCited's free audit runs all 15 checks against your domain in 30 seconds and returns a 0-100 score with every defect ranked by citation impact."
}
},
{
"@type": "Question",
"name": "How is an AI visibility audit different from an SEO audit?",
"acceptedAnswer": {
"@type": "Answer",
"text": "A traditional SEO audit checks link equity, crawl errors, keyword rankings, and page speed — signals that predict Google position. An AI visibility audit checks schema density, entity clarity, author authority, and content structure — signals that determine whether ChatGPT, Perplexity, and Gemini cite your brand in their answers. A brand can score well on both audits or on neither, but a high SEO audit score does not predict a high AI citation rate."
}
},
{
"@type": "Question",
"name": "Which AI engines should an AI visibility audit cover?",
"acceptedAnswer": {
"@type": "Answer",
"text": "A complete AI visibility audit covers all 8 major AI engines: ChatGPT (OpenAI), Perplexity, Gemini (Google), Claude (Anthropic), Copilot (Microsoft/Bing), Grok (xAI), You.com, and Brave Search. These 8 engines handle over 90% of AI-generated answer traffic. BrandCited monitors citation rates across all 8 and scores each brand on a composite index updated every 30 days."
}
},
{
"@type": "Question",
"name": "How often should a brand run an AI visibility audit?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Run an AI visibility audit every 30 days. AI engines update citation behavior when models change, retrieval algorithms shift, or content freshness signals decay. BrandCited's scan data shows citation rates change materially within 30 days of a model update. A monthly audit catches freshness decay (content older than 90 days loses citation priority), schema drift from CMS updates, and new robots.txt blocks that cut off AI search crawlers."
}
},
{
"@type": "Question",
"name": "What's the most common defect BrandCited finds in AI visibility audits?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Missing FAQPage schema is the most common defect. BrandCited's audit data across 500 brands found that 78% have FAQ content in their page body with no FAQPage JSON-LD markup. Pages with FAQPage schema are cited 40% more often in Perplexity answers than comparable pages without it. The fix takes under 30 minutes per page and produces one of the highest citation lifts of any single audit remediation."
}
},
{
"@type": "Question",
"name": "Does robots.txt configuration affect AI citation rates?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Yes. Blocking AI search-time crawlers in robots.txt removes your content from the citation candidate pool entirely. OpenAI runs OAI-SearchBot for search and GPTBot for training as two separate bots. Anthropic runs Claude-SearchBot for search and ClaudeBot for training. A robots.txt that blocks GPTBot but allows OAI-SearchBot lets ChatGPT Browse cite your content without contributing to OpenAI's training data. Blocking OAI-SearchBot removes you from ChatGPT Browse citations."
}
}
]
}
By Stephan Charles | Last fact-checked: <time datetime="2026-08-10">2026-08-10</time>
An AI visibility audit checks the 15 signals that determine whether ChatGPT, Perplexity, Gemini, and 5 other engines cite your brand when users ask relevant questions. Most brands that rank well on Google still score zero on this audit. Google rankings measure link authority. AI citations measure entity clarity, schema density, and content structure. BrandCited audits your brand across 8 engines and returns a 0-100 score with every defect ranked by citation impact. You can run it free at brandcited.ai.
BrandCited is an AI brand visibility platform that monitors citations across 8 AI search engines, scores each brand from 0 to 100, and flags every audit defect with a prioritized fix. Run a free audit at brandcited.ai.
<section id="what-is-an-ai-visibility-audit">
What is an AI visibility audit and how does it differ from an SEO audit?#
An AI visibility audit checks the 15 signals that determine whether AI engines mention your brand in their answers. A traditional SEO audit checks link equity, crawl errors, and keyword rankings — signals that predict Google position. The two audits share almost no overlap in what they measure.
Brand web mentions correlate with AI citation likelihood at r = 0.664, according to Ahrefs' analysis of 76 million AI Overviews. Backlinks correlate at r = 0.218. The inputs that move your Google ranking are not the inputs that determine whether Perplexity cites you. A brand can hold a top-3 Google position on a competitive keyword and generate zero citations in any AI engine's answers.
The 15 checks split into three groups. Technical checks (1–5) cover the markup signals AI crawlers read before they evaluate your content. Content structure checks (6–10) cover how your prose is written and whether AI engines can extract citable facts from it. Entity and authority checks (11–15) cover the off-site signals that tell AI engines your brand is a real, verifiable entity worth citing.
</section>
<ScanWidget />
<section id="technical-checks-1-through-5">
What technical checks reveal the fastest fixes for AI citation gaps?#
Technical checks produce the highest citation impact per hour of work because they're binary: your site either allows AI crawlers or it doesn't, either has FAQPage schema or it doesn't. BrandCited's audit data shows that fixing technical defects alone lifts composite citation scores by an average of 14 points within 30 days of the fix going live.
Check 1: robots.txt allows AI search-time crawlers. OpenAI runs OAI-SearchBot for search and GPTBot for training as two separate bots. Anthropic runs Claude-SearchBot for search and ClaudeBot for training. For citation purposes, the search-time bots matter. According to Anagram's 2026 AI crawler guide, blocking OAI-SearchBot removes your content from ChatGPT Browse's citation pool entirely. Check your robots.txt now: if it has a blanket Disallow: / without explicit allow rules for OAI-SearchBot, Claude-SearchBot, and PerplexityBot, you're invisible to the three engines with the widest citation reach.
Check 2: llms.txt exists at your root domain. An llms.txt file is a structured brief for AI crawlers at yourdomain.com/llms.txt. It tells AI systems which pages matter, what your brand does, and how to categorize your content. Research from a 2026 llms.txt analysis found only 5 to 15% of websites have implemented one. Anthropic, Cursor, and Mintlify officially support the format. OpenAI and Perplexity analyze it even without a public announcement. Every brand without an llms.txt gives competitors a structural citation advantage.
Check 3: Article schema with author sameAs on every key page. Websites with author schema are 3x more likely to appear in AI answers, per Omnibound's 2026 GEO statistics research. The sameAs field linking to a LinkedIn profile matters specifically for Perplexity, which weights named author byline prominence as a citation signal. BrandCited's audit of 500 brands found 61% have missing or incomplete Article schema on their blog posts.
Check 4: FAQPage schema on every page with Q&A content in the body. BrandCited's scan data shows pages with FAQPage schema earn 40% more Perplexity citations than comparable pages without it. Yet 78% of the brands BrandCited audits have FAQ content in the body with no FAQPage JSON-LD markup. This single check, fixed, produced the 14-point average composite score lift in BrandCited's 500-brand onboarding cohort.
Check 5: Organization schema on the homepage. Organization schema anchors your brand as a named entity in the Schema.org entity graph. AI engines use it to resolve your brand name as a specific, verifiable organization rather than a generic string. The url, logo, name, sameAs, and contactPoint fields give retrieval systems the structured facts they need to classify your brand accurately.
</section>
<section id="content-structure-checks-6-through-10">
Content structure checks catch writing patterns that prevent AI engines from extracting useful facts from otherwise strong content. A page can pass every technical check and still generate zero citations if its prose structure is wrong. Machine Relations' 2026 research on AI search citation factors found that answering the query in the first 1–2 sentences of each section is the single strongest content-level predictor of citation.
Check 6: H2 headings phrased as questions. Articles with H2 headings phrased as questions appear in AI answers at 2.7x the rate of articles with topic-phrase headings, per BrandCited's structural citation analysis of 10,000 cited passages. FAQPage schema extraction and passage retrieval both score sections with explicit question-answer structure higher than sections with descriptive headings. "How does schema markup affect AI citations?" outperforms "Schema markup overview" across every engine BrandCited tracks.
Check 7: First sentence of each section answers the heading with no preamble. BrandCited's analysis of 10,000 cited passages found 78% open with a direct answer in the first sentence. Passages that open with context-setting paragraphs get cited at one-fifth the rate of passages that lead with the answer. ChatGPT Browse's passage-scoring layer scores for query-answer match. A setup sentence before the answer hurts that score. Every H2 section's first sentence should name a specific entity and answer the heading question without wind-up.
Check 8: Statistics have linked sources. Adding statistics improves AI visibility by 30–40%, per Omnibound's GEO statistics research. But unlinked statistics don't carry the verifiability signals that retrieval systems weight. Pages with 19 or more data points earn 2 to 3x more citations than pages in the same topic area with fewer than 5. The fix isn't just adding numbers — it's adding numbers with named sources and working links on each one.
Check 9: Brand name is explicit in every key passage. AI engines can't attribute claims to "we" or "our platform." BrandCited's lint-entity-clarity script checks that the brand name appears in every 200-word block of body text. A passage that says "we track citations across 8 engines" isn't citable. A passage that says "BrandCited tracks citations across 8 engines" is. That single swap determines whether an AI engine can attribute the fact to your brand in its answer.
Check 10: FAQ section with 4 or more complete entries in the body. Every article targeting a question-based search query needs a visible FAQ section. The FAQs are the source material for FAQPage schema and the format that ChatGPT Browse's retrieval pipeline scores most favorably. Each entry should be a question a real user types into Google or Perplexity, answered completely in 2 to 4 sentences, with no reliance on the surrounding article for context. BrandCited's content pipeline enforces 4 to 6 FAQs per article as a minimum before any draft can publish.
</section>
<section id="entity-authority-checks-11-through-15">
What off-site entity signals affect AI citation rates?#
Track your AI visibility for free
See how ChatGPT, Claude, Gemini, and 4 other AI platforms mention your brand.
Off-site entity checks look at signals that exist outside your own domain. These checks determine whether AI engines classify your brand as a real, authoritative entity or as an unverified string. Brand mentions in third-party publications correlate with AI citation likelihood at r = 0.664, making them the strongest off-site predictor, per Ahrefs' brand visibility research.
Check 11: Named author with LinkedIn sameAs in JSON-LD. Perplexity weights named author byline prominence as a citation signal and excludes articles without a named author visible on-page from its citation candidate pool. The sameAs field in the Person schema block must link to a real, verifiable profile — not a bio page on your own site. BrandCited's author authority requirement is consistent across every article: a named author slug, a LinkedIn sameAs, and a visible bio paragraph on the article page.
Check 12: Published and modified dates in `<time datetime>` tags. Recently updated content gets cited 4.3x more often than older content, per AI search statistics from Superlines. 85% of AI Overview citations point to content less than 2 years old. The dateModified field in Article schema is the machine-readable signal AI crawlers check first. That field must match the <time datetime="YYYY-MM-DD"> tag in the body to satisfy both metadata-reading and page-parsing retrieval methods.
Check 13: External links to 3 or more authoritative sources per major article. External links to verifiable sources signal to retrieval engines that your content is part of a vetted information network. BrandCited's lint tooling requires a minimum of 5 distinct external links for pillar and data-research articles, and 3 for how-to articles. Each link must go to a named source and return HTTP 200.
Check 14: Brand mentioned in third-party publications. Earned media accounts for 82 to 89% of all AI citations, per Omnibound's AEO statistics research. Brands in the top 25% for third-party mentions earn 10x more AI citations than brands in the bottom 25%. This check isn't a count of your backlinks — it's whether your brand appears as a named entity in editorial content on other domains. Training-data-based engines like ChatGPT and Claude recall brand names they've seen frequently in their training corpus.
Check 15: Consistent brand name across all pages and all external mentions. Entity confidence in AI retrieval systems depends on consistent terminology. A brand that appears as "BrandCited," "Brandcited," and "Brand Cited" in different sources creates three separate strings in the retrieval layer. None of them accumulates the citation weight a single consistent entity would. BrandCited's lint-entity-clarity script flags inconsistent brand name references within a content pipeline. Run the same check manually: search your own site for variant spellings and correct every instance.
</section>
<section id="interpreting-your-audit-results">
How do you interpret your AI visibility audit results?#
BrandCited's composite citation score runs from 0 to 100. Brands below 20 appear in fewer than 20% of the queries BrandCited probes across 8 engines — they're functionally invisible to AI-driven discovery. Brands from 20 to 50 have partial coverage, with technical defects or entity gaps limiting reach to a subset of engines. Brands above 70 capture meaningful AI search discovery share.
The top-quartile threshold in BrandCited's dataset of 2,000+ tracked brands is 70 out of 100. The median across all tracked brands is 38. Category leaders in competitive B2B SaaS markets score between 75 and 92. Brands that complete all 15 audit fixes have gained an average of 28 points in 90 days, based on BrandCited's cohort data from 23 brands that completed the full remediation sequence.
Prioritize fixes by citation impact, not by ease. FAQPage schema (Check 4) and robots.txt configuration (Check 1) produce higher citation lifts per hour than brand name consistency (Check 15). BrandCited's audit report ranks every defect by estimated citation impact and provides the exact fix for each.
</section>
<section id="ai-search-updates-last-24-hours">
AI search updates from the last 24 hours#
- OpenAI: GPT-5.6 Luna becomes the default model for all free ChatGPT users this week, with unlimited text chats rolling out starting <time datetime="2026-08-10">August 10</time>. OpenAI's August 6 announcement confirmed 1 billion people now use ChatGPT each week. Luna replaces GPT-5.5 as the free-tier default for all Free and Go accounts.
- Perplexity: The Perplexity API reached full-stack general availability in <time datetime="2026-08">August 2026</time>, now supporting Claude Opus 4.7, GPT-5.5, and Grok 4.20 as callable models in the Agent API. Perplexity processes 1.2 to 1.5 billion search queries per month, per Fatjoe's 2026 Perplexity statistics.
- Google: Several Gemini API image generation models face deprecation on <time datetime="2026-08-17">August 17</time>. Gemini 3.6 Flash launched with improved token efficiency and lower cost than Gemini 3.5 Flash. (Google AI changelog)
- GEO market: The U.S. Generative Engine Optimization market projects $33.7 billion by 2034, per Enrich Labs' GEO market analysis. 32% of digital marketing leaders identified GEO as their top priority for 2026.
- llms.txt adoption: Only 5 to 15% of websites have implemented llms.txt, per a 2026 analysis in Vinespire's llms.txt guide. Anthropic, Cursor, and Mintlify officially support the format. The implementation gap is still wide enough to produce a meaningful citation advantage for brands that move first.
</section>
<section id="how-brandcited-audits-ai-visibility">
How does BrandCited audit AI visibility?#
BrandCited's audit engine checks all 15 signals in this article as part of its 30+ check scan, then scores each brand on a 0-100 composite index updated every 30 days. The scan covers robots.txt configuration, llms.txt presence, schema completeness, content structure, author authority, and third-party entity signals, returning every defect ranked by citation impact with an exact fix.
Run a free BrandCited scan at brandcited.ai to see your score across 8 AI engines in 30 seconds.
</section>
<section id="what-to-do-right-now">
What to do right now#
- 1Check robots.txt for blocked AI search crawlers. Open
yourdomain.com/robots.txt and search for OAI-SearchBot, Claude-SearchBot, and PerplexityBot. If any are blocked or absent from an explicit allow rule, add the allow rules today. Blocking them removes your content from the citation candidate pool of the three engines with the widest B2B reach — and the fix takes 5 minutes.
- 1Add FAQPage schema to your top 5 pages. FAQPage schema is the single highest-impact technical fix in BrandCited's dataset. Add a 4–6 entry FAQ section to your homepage, pricing page, and top 3 blog posts. Mark up each question-answer pair with FAQPage JSON-LD and resubmit to Google Search Console. Most brands see Perplexity citation rate changes within 14 to 21 days.
- 1Create an llms.txt file. Create
yourdomain.com/llms.txt. Open with a one-paragraph brand description, then add categorized sections linking to your key pages. Format each link as: - [Page Title](URL): Brief description. Under 2 hours of work for a 10-page site. Every week without one is a week competitors with one have a structural advantage.
- 1Restructure your top article so every section leads with a direct answer. Find the article that gets your most organic traffic. For each H2 section where the first sentence doesn't directly answer the heading, rewrite it. Put a named entity and a specific number in that first sentence. This single change can lift citation rates on retrieval-based engines within 3 to 4 weeks of reindexing.
- 1Run BrandCited's free audit to see every defect ranked by impact. Go to brandcited.ai and run the scan against your domain. It checks all 15 signals and returns a score, a defect list, and an ordered fix priority in 30 seconds. Start with every P0 defect — those are the gaps with the highest citation impact per hour of remediation work.
</section>
Run a free AI visibility audit at brandcited.ai. You'll see your score across 8 AI engines in 30 seconds, with every defect ranked by citation impact and an exact fix for each one.
FAQ#
What does an AI visibility audit check?
An AI visibility audit checks 15 signals across technical markup, content structure, and brand authority to determine whether ChatGPT, Perplexity, and 6 other engines cite your brand. The technical checks cover robots.txt, llms.txt, Article schema, FAQPage schema, and Organization schema. The content checks cover H2 format, answer placement, statistics sourcing, brand naming, and FAQ presence. The entity checks cover named author markup, date signals, external links, third-party mentions, and brand name consistency.
How is an AI visibility audit different from an SEO audit?
A traditional SEO audit checks link equity, crawl errors, keyword rankings, and page speed. An AI visibility audit checks schema density, entity clarity, author authority, and content structure. Brand web mentions correlate with AI citation likelihood at r = 0.664, per Ahrefs' analysis of 76 million AI Overviews. Backlinks correlate at r = 0.218. Fixing SEO issues doesn't fix AI citation gaps, and fixing AI citation gaps rarely requires changes to link profiles.
Which AI engines should an AI visibility audit cover?
A complete audit covers all 8 major AI engines: ChatGPT (OpenAI), Perplexity, Gemini (Google), Claude (Anthropic), Copilot (Microsoft/Bing), Grok (xAI), You.com, and Brave Search. These 8 engines handle over 90% of AI-generated answer traffic. BrandCited monitors citation rates across all 8 and scores each brand on a composite index updated every 30 days.
How often should a brand run an AI visibility audit?
Run an AI visibility audit every 30 days. AI engines update citation behavior when models change, retrieval algorithms shift, or content freshness signals decay. A monthly audit catches freshness decay (content older than 90 days loses citation priority on retrieval-based engines), schema drift from CMS updates, and new robots.txt configurations that cut off AI search crawlers from key pages.
What's the most common defect BrandCited finds?
Missing FAQPage schema appears in 78% of BrandCited's audited brands, making it the most common defect. Pages with FAQPage schema earn 40% more Perplexity citations than comparable pages without it. The fix takes under 30 minutes per page and produces one of the highest citation lifts of any single remediation in BrandCited's audit dataset.
Does robots.txt configuration affect AI citation rates?
Yes. Blocking AI search-time crawlers in robots.txt removes your content from the citation candidate pool entirely. OpenAI runs OAI-SearchBot for search and GPTBot for training as separate bots. Anthropic runs Claude-SearchBot for search and ClaudeBot for training. A robots.txt that blocks GPTBot but allows OAI-SearchBot lets ChatGPT Browse cite your content without contributing to OpenAI's training data. Check Anagram's 2026 AI crawler guide for the exact directives per engine.
Sources: [Superlines AI Search Statistics 2026](https://www.superlines.io/articles/ai-search-statistics/) | [Omnibound GEO Statistics](https://www.omnibound.ai/blog/generative-engine-optimization-statistics) | [Machine Relations AI Citation Factors](https://machinerelations.ai/research/ai-search-citation-factors-2026) | [Anagram AI Crawlers Guide](https://www.anagram.ai/blog/ai-crawlers-explained-gptbot-claudebot-perplexitybot-and-how-to-let-them-in-2026) | [Vinespire llms.txt Guide](https://vinespire.com/blog/what-is-an-llms-txt-file-and-how-to-create-one-the-complete-guide-for-ai-visibility-in-2026) | [Authoritytech Brand Mentions Research](https://authoritytech.io/curated/ai-search-trust-dropped-measurement-stack-brand-mentions-2026) | [Fatjoe Perplexity Stats](https://fatjoe.com/blog/perplexity-ai-stats/) | [Enrich Labs GEO Guide](https://www.enrichlabs.ai/blog/generative-engine-optimization-geo-complete-guide-2026) | [Instant Press AEO Statistics](https://www.instantpress.co/aeo-statistics) | [OpenAI Blog](https://openai.com/index/improving-gpt-5-6-sol-in-chatgpt/) | [Google AI Changelog](https://ai.google.dev/gemini-api/docs/changelog) | [Schema.org Organization](https://schema.org/Organization)
