How to Run an AI Visibility Audit: A Step-by-Step Guide

An AI visibility audit is the structured assessment that tells you exactly how AI answer engines currently see, cite, and characterise your brand, and where the gaps are. Without one, generative engine optimisation is guesswork: you are publishing content and hoping it helps, with no baseline and no diagnosis. An audit replaces that with a clear picture of your current standing, your competitors' standing, and a prioritised list of what to fix first.

This guide walks through a repeatable audit you can run yourself, in roughly the order the steps depend on each other. It starts with the technical foundation (can the engines even access you), moves through current citation reality and competitive position, then into the content and entity fixes that move the needle, and ends with how to report it so the work gets funded. Run it as a baseline now, then re-run it quarterly to measure progress.

Step 1: Check crawler and technical access

Nothing else matters if the engines cannot read your site, so start here.

Verify AI crawler access. Check your robots.txt and any bot-management or firewall rules for whether you are allowing or blocking the AI crawlers (GPTBot, OAI-SearchBot, PerplexityBot, Google-Extended, ClaudeBot, and others). Many sites block these by accident or by stale default. Decide deliberately which to allow; blocking them makes you invisible to that engine.

Confirm content is server-rendered or crawlable. A JavaScript single-page app that returns an empty shell to a non-executing crawler is invisible regardless of content quality. Verify that your key pages deliver real content in the initial HTML response, not only after client-side rendering.

Check structured data and basics. Confirm your Organization, Product, and other relevant schema is present and valid, that pages have clean titles and headings, and that your site is technically sound. These fundamentals are covered in how to optimise for ai search.

Step 2: Establish your current citations and baseline

Now measure where you actually stand. Run your prompt set across the engines your buyers use and record, for each prompt, whether you are cited, how prominently (named first, mentioned, or absent), and how the engine characterises you (accurate, vague, or wrong). If you have not built a prompt set yet, do that first; our guide to prompts to track shows how. This baseline is the number every later measurement is compared against, so record it carefully. A tool like bing.ly automates running a prompt set across multiple engines and capturing citation presence and sentiment, which makes the baseline repeatable rather than a one-off manual slog.

Step 3: Measure share of voice and the competitor gap

A citation rate means little without context, so measure it against competitors.

Calculate share of voice. Across your prompt set, work out how often you are cited versus your named competitors. This share-of-voice figure is the single most useful summary metric, because it tells you whether you are winning or losing the AI conversation in your category, not just whether you appear at all.

Identify the competitor gap. For the prompts where competitors are cited and you are not, note who wins and why: do they have stronger structured data, better comparison content, more reviews, a clearer entity, editorial coverage you lack? This gap analysis is where your action plan comes from, because it points at the specific advantages you need to close.

Note characterisation gaps. Where the engine describes you inaccurately or vaguely, that is a content and entity clarity problem to fix, separate from pure absence.

Step 4: Diagnose the content and entity fixes

Translate the gaps into fixes, prioritised by impact.

Content gaps. Missing category pages, comparison and alternative content, use-case pages, and answer-first formatting. Where competitors are cited for a question you have no good page for, that is a content gap to fill.

Entity gaps. Inconsistent naming or NAP, missing Organization schema, no Wikidata presence, no sameAs links connecting your profiles. Entity clarity underpins citation confidence; our entity seo for ai guide covers the fixes.

Authority gaps. Thin reviews, no presence in the editorial roundups the engine cites, weak third-party corroboration. These take longer but matter for competitive queries.

Prioritise fixes by impact and effort: high-impact, low-effort fixes (a missing comparison page, broken crawler access, absent schema) first.

Step 5: Report it so the work gets funded

An audit nobody acts on is wasted. Report it as: current citation rate and share of voice, the competitor gap with named examples, the prioritised fix list with expected impact, and a baseline to measure against. Frame it in business terms (you are absent from the answer your buyers read while competitor X is recommended) rather than vanity metrics. For setting honest expectations on what GEO can and cannot deliver, pair the report with does geo work, and connect it to traffic via ai citation tracking. Then re-run the full audit quarterly to show movement.

Frequently Asked Questions

Q: What is the first thing to check in an AI visibility audit? Crawler and technical access. Confirm you are not blocking AI crawlers in robots.txt or your firewall, and that your key pages deliver real content in the initial HTML rather than only after client-side rendering. If the engines cannot read you, nothing else in the audit matters.

Q: How do I measure share of voice? Run your prompt set across the relevant engines and calculate how often you are cited versus your named competitors across that set. The resulting percentage is the clearest summary of whether you are winning or losing the AI conversation in your category.

Q: How long does an AI visibility audit take? The technical and baseline measurement can be done in days, especially with a tool to automate running the prompt set across engines. The content, entity, and authority fixes that follow are an ongoing programme. Treat the audit as a recurring quarterly baseline, not a one-time project.

Q: How often should I re-run the audit? Quarterly. AI engines change, competitors move, and your own content evolves, so a quarterly re-run shows whether your fixes improved citation rate and share of voice. Keep your prompt set stable across runs so the comparison is meaningful.

The Bottom Line

An AI visibility audit turns generative engine optimisation from guesswork into a diagnosed, prioritised programme. Start with crawler and technical access, baseline your current citations and characterisation, measure share of voice and the competitor gap, then translate the gaps into prioritised content, entity, and authority fixes. Report it in business terms so the work gets funded, automate the measurement so it is repeatable, and re-run it quarterly to prove progress. The brands that audit honestly and act on the gaps are the ones that climb the AI answers while their competitors guess.