The Complete AI SEO Audit: How to Find and Fix Every Gap

TL;DR: An AI SEO audit checks five layers — technical crawlability, schema markup, entity recognition, content structure, and cross-platform citation presence. This guide walks through each layer in order, with exact checks to run and fixes to apply. Budget 3–4 hours for a thorough self-audit.

Why a dedicated AI SEO audit is different from a traditional audit

A standard SEO audit checks keywords, backlinks, site speed, and crawlability. An AI SEO audit checks an almost entirely different set of signals — the ones that determine whether ChatGPT, Perplexity, Google AI Mode, and Gemini cite your brand when buyers ask questions in your category.

The critical data point: 80% of URLs cited by AI systems do not rank in Google's top 100 for the same query. This means your traditional SEO audit is largely blind to your AI visibility gaps. You need a separate, systematic audit focused on AI-specific signals.

Layer 1: Technical AI crawlability audit

Before any content or schema optimisation can work, AI crawlers need to be able to access, index, and parse your site. Check each of the following:

• Bing Webmaster Tools — is your site verified and is your sitemap submitted? ChatGPT Browse runs on Bing. If you are not indexed by Bing, you are invisible to ChatGPT for real-time queries. Verify at bing.com/webmasters.
• robots.txt — are AI crawlers explicitly allowed? Check that GPTBot, PerplexityBot, ClaudeBot, and Google-Extended are not blocked. A single disallow line can eliminate your AI visibility entirely.
• llms.txt — does your site have an llms.txt file at the domain root? This plain-text file tells AI systems what your site is about and which pages to prioritise. It is the AI equivalent of a sitemap. See our complete llms.txt guide.
• Page speed — AI crawlers behave like browsers. Sites with Core Web Vitals failures get deprioritised. Run PageSpeed Insights on your 5 most important pages.

Layer 2: Schema markup audit

Schema markup is the most direct technical signal for AI citation. For each key page, check:

• Does it have the correct schema type? (Person, Organization, Service, Article, FAQPage, HowTo)
• Are there any validation errors? Use Google Rich Results Test on every page.
• Does your Organization schema include a sameAs array linking to LinkedIn, Crunchbase, and Wikidata?
• Does each service page have FAQPage schema with 5+ questions phrased as users would ask them?
• Are your blog posts tagged with Article schema including author, datePublished, and dateModified?

Missing or broken schema on even one key page significantly reduces your citation rate. Our Schema Markup service covers the complete implementation. See also our schema markup technical guide.

Layer 3: Entity recognition audit

AI systems cite brands they "know" — entities they can verify across multiple authoritative sources. Check your entity footprint:

• LinkedIn — complete company page with consistent name, description, and URL? LinkedIn is the most-cited domain for professional queries across all major AI platforms.
• Crunchbase and G2 — are you listed? G2 is the most-cited software review platform on ChatGPT and Perplexity.
• Wikidata — do you have a Wikidata entry? This is one of the strongest entity verification signals for AI systems.
• NAP consistency — is your Name, Address, Phone identical across all directories?
• Person schema sameAs — does your Person schema link to all your verified profiles?

Layer 4: Content structure audit

Check each key page for AI-extraction-ready structure:

• Does it open with a TL;DR summary? (44.2% of all LLM citations come from the first 30% of text)
• Are subheadings phrased as questions users would ask AI?
• Are paragraphs 3–4 sentences maximum?
• Does it contain specific statistics with sources? (Pages with stats get 30–40% more AI citations)
• Is there a FAQ section at the bottom with FAQPage schema?

Layer 5: Citation presence audit

The final layer is measuring your actual citation rate — how often your brand appears when real users ask relevant questions.

• Open ChatGPT and run 10–15 queries your ideal buyer would ask. Is your brand mentioned?
• Do the same on Perplexity. (Perplexity cites sources in 97% of responses vs ChatGPT's 16% — Perplexity is easier to track.)
• Check Google AI Mode for your target queries.
• Document which competitors are being cited instead of you — this tells you which gaps to prioritise.

This manual audit gives you a baseline citation rate. For ongoing monitoring, see our AI Citation Monitoring service.