You cannot optimize what you have not measured. Yet most companies investing in AI visibility have never conducted a structured audit of their current standing. They publish content, implement schemas, and hope for citations, without ever establishing a baseline, identifying their specific gaps, or prioritizing actions by impact.
An AI visibility audit is the systematic process of measuring where you are currently cited, where you should be cited and are not, why you are being ignored, and what specific actions will move the needle fastest. Done correctly, it transforms AI visibility from a vague aspiration into a clear, prioritized action plan.
This guide presents the complete audit methodology we use at AISOS. It covers technical auditing (crawlability, Schema, page structure), content auditing (query coverage, format, freshness), authority auditing (entity presence, external mentions), and monitoring setup (how to track changes over time). Whether you are starting from zero or fine-tuning an existing strategy, this methodology applies. To understand the scoring system this audit produces, see our AI Visibility Score guide.
Phase 1: Technical access audit
The technical access audit answers one question: can AI systems reach and read your content at all? A single misconfiguration here makes every subsequent optimization effort irrelevant. Start with your robots.txt file. Fetch yoursite.com/robots.txt directly and check for any Disallow rules affecting GPTBot, Google-Extended, ClaudeBot, or PerplexityBot. Then check HTTP headers: some servers send X-Robots-Tag: noindex headers that block AI crawlers even when robots.txt is permissive. Use a tool like curl with a user-agent string matching each crawler to verify what they actually receive.
Next, audit your page rendering. Open 10 representative pages and view the HTML source (not the rendered DOM). If your main content does not appear in the initial HTML, AI crawlers are not seeing it. This is a common issue with JavaScript-heavy frameworks that rely entirely on client-side rendering. Server-side rendering or static generation resolves it. Also verify page speed: AI crawlers have crawl budget limits per domain, and slow pages consume more budget, leading to incomplete crawling of your content library.
Complete the technical access audit with a review of your XML sitemap. Verify that the sitemap includes all important content pages, that lastmod dates reflect genuine update dates (not fictitious timestamps), and that the sitemap URL is correctly declared in robots.txt. AI crawlers use the sitemap to prioritize crawling order. A well-maintained sitemap with accurate dates ensures your most recently updated and most important content is crawled first. This is foundational technical SEO that matters doubly for AI crawlers.
Phase 2: Schema and structured data audit
Schema markup is how your site communicates its structure and authority to AI systems. Your Schema audit should cover four questions: what schemas are currently implemented, are they implemented correctly, are they implemented comprehensively, and are they the right schemas for your content types?
Extract all Schema markup from your key pages using a crawling tool or the Schema Markup Validator. For each page type (homepage, product pages, blog posts, FAQ pages, team pages), verify that the appropriate schema is present and that required properties are populated. Common gaps include: Article schemas missing author or dateModified, Organization schemas missing sameAs properties, FAQPage schemas on pages with Q&A content but no schema declaration, and product pages with no Product schema despite having pricing and features.
Severity rank your Schema gaps by page type and traffic volume. Schema errors on your highest-traffic pages cost you the most in citation potential. Fix Organization and Person schemas first (they affect every page through entity graph relationships), then FAQPage and HowTo (they directly drive citation events), then Article schemas on your most important content pages. The advanced Schema guide details the implementation specifics for each type. After fixing, revalidate with both Google Rich Results Test and schema.org validator to catch different error categories.
Phase 3: Query coverage and citation testing
This is the core of the audit: testing whether AI systems actually cite you when your prospects ask relevant questions. Build a query set of 25 to 40 questions representing the decisions your customers make before buying from you. These should be phrased as natural language questions, not keywords. "What is the best HR software for a 200-person manufacturing company?" not "best HR software manufacturing."
Test each query across ChatGPT (GPT-4o), Perplexity, and Gemini in clean private browsing sessions to avoid personalization bias. For each response, record: whether your brand appears, whether a specific page URL is cited, whether the reference is positive, neutral, or comparative, and which competitors appear when you do not. This creates your citation map, showing exactly where you are visible and where you are invisible.
Calculate your citation rate (citations divided by total query-LLM combinations) and segment by query category. Most companies discover that their citation rate varies dramatically by topic even within their niche: they may be cited reliably for product feature queries but invisible for problem-awareness queries, or vice versa. These gaps identify exactly where to focus content creation effort. Compare your citation rate against the AI Visibility Score benchmarks to understand where you stand relative to sector averages and what score puts you in the top quartile for your industry. Also note competitor citations: if the same three competitors appear in every gap, analyze their cited pages to understand what they are doing that you are not.
Phase 4: Content quality and format audit
Once you know which queries you are not winning, the content audit explains why. Pull the pages that should theoretically be ranking for your gap queries and evaluate them against five criteria: directness (does the first paragraph directly answer the question?), structure (does the page use hierarchical headers, lists, and tables?), originality (does the page contain data, analysis, or perspective that ChatGPT could not generate itself?), freshness (is the content dated and updated within the last six months?), and depth (does the page cover the topic comprehensively enough to be a reference source?).
Score each page one to five on each criterion. Pages scoring below three on any criterion are candidates for optimization. Pages scoring below three on multiple criteria are the primary drag on your citation rate. Prioritize optimization by combining the score with the business importance of the query: a page covering a high-intent purchase query that scores two on directness and one on originality is your highest-priority fix, regardless of how long the page has existed or how much effort went into creating it.
Also audit your content gap: queries where you have no page at all. These represent content creation opportunities with a defined target format. For each gap query, identify what format the cited competitor pages use (HowTo guide, comparison table, FAQ page, case study with data) and use that as your template for the new page. The content clustering guide explains how to organize these new pages for maximum topical authority impact.
Phase 5: Entity presence and external mention audit
AI systems build entity graphs from the sources they train and retrieve from. Your entity presence audit checks whether these sources know you exist and describe you accurately. Start with the most important signals: Wikipedia eligibility (do you meet notability criteria?), presence in industry-specific directories and databases, mentions in trade media and reference publications, and citations in educational or research content in your field.
Search for your brand name in the sources LLMs are known to weight heavily: trade publications in your vertical, professional directories (G2, Capterra, Clutch for software), industry associations, and high-authority blogs in your niche. Document every mention, noting the source domain authority, the context of the mention (positive, neutral, comparative), and whether it includes a link to your site. This is your external mention baseline.
Gaps in entity presence are often the hidden cause of low citation rates even when on-site optimization is strong. A company with perfect Schema markup and excellent structured content may still be invisible to AI systems that have no external evidence of the company's existence or authority. The fix requires active outreach: guest contributions to industry media, participation in curated directories, and citation in third-party content. For a sector-specific view of how entity presence drives citations, see our consulting industry AI visibility guide.
Building your audit-to-action roadmap
An audit without a prioritized action plan is just documentation. The final phase of the AI visibility audit is converting findings into a sequenced roadmap. Sort your identified issues into three buckets: technical blockers (issues that prevent AI from accessing or understanding your content at all, fix these first regardless of effort), high-impact content gaps (queries with business importance where you have no or poor content, create or fix these second), and authority gaps (missing entity presence that suppresses citations despite good on-site content, build these third through sustained outreach).
Within each bucket, prioritize by the ratio of impact to effort. Fixing a robots.txt issue that is blocking GPTBot takes five minutes and can immediately increase citation rate across all your content. Restructuring a high-traffic page for answer-first format takes half a day and can convert an invisible page into a frequently cited source. Publishing a new piece of content on a gap query takes a week but opens an entirely new citation opportunity. Stack these actions into a 90-day sprint with clear weekly milestones.
Set up monitoring before you start implementing changes. Without a pre-implementation baseline captured during the audit, you will not be able to measure what your actions actually achieved. AISOS automates this monitoring with monthly multi-LLM reports, but even a manual monthly test across 25 queries on three LLMs gives you the signal you need to know whether your roadmap is working. The AI Visibility Score guide explains exactly what to measure and how to interpret monthly score changes.