GEO Audit: Measure Your Brand's Visibility in AI Engines

Search is splitting in two. Alongside Google's familiar list of blue links, a second surface has taken hold: the written answer from an assistant like ChatGPT, which synthesises a response and then cites a handful of sources. For a business, the question is no longer only "do I rank well?" but "am I the brand the AI names when it answers my customers?". A GEO audit is the tool that answers that second question, with numbers rather than a hunch. Here is what it looks at, how it scores, and then how to read what it sends back.

What a GEO audit is

A GEO audit (Generative Engine Optimization) measures your brand's visibility in the answers of AI engines such as ChatGPT and Perplexity. Starting from your real business queries, it establishes whether you are cited, then analyses why.

The term GEO comes from a foundational piece of research. In 2024, a team including researchers from Princeton and the Allen Institute for AI published the study GEO: Generative Engine Optimization, presented at the ACM SIGKDD conference, which laid out for the first time a rigorous framework for optimising a piece of content's visibility in generated answers. A GEO audit is the practical application of that discipline: before optimising, you measure where you stand.

It is worth clearly separating a GEO audit from a classic SEO audit. An SEO audit tells you whether you can rank in Google's list of links. A GEO audit tells you whether you can be cited in an answer written by an AI. These are two different surfaces, two different reward systems, and therefore two different scoring grids, even though they share common foundations we detail further down.

A GEO audit done well does not produce an abstract score of the "you are at 62 out of 100" kind. It produces two concrete things: a quantified snapshot of your current presence in AI answers, recorded query by query, plus a ranked list of actions running from quick fixes to deeper work. It is this orientation toward action that separates a real audit from a mere observation.

Why audit your AI visibility now

Auditing your AI visibility is worthwhile right now because your customers already use ChatGPT and the other assistants to inform themselves and compare options, because citation by these engines brings visibility that most of your competitors do not yet measure, and because the work puts you ahead before any wider rollout of Google's AI Overviews.

One important point about markets where the rollout is uneven. To this day, Google's AI Overviews, those AI summaries shown at the top of the results, are not officially deployed everywhere, sometimes for regulatory reasons tied to press neighbouring rights. You might conclude there is no urgency. That would be a misreading, for two reasons.

The habit has already shifted: putting your question to an assistant before opening a search engine.

First reason: ChatGPT and Perplexity are already heavily used by the public, web search included, regardless of AI Overviews. ChatGPT's web search, launched in late 2024 and then opened to all logged-in users, works the world over. Your prospects already ask it their comparison and decision questions there. Visibility in those answers is decided today, not on the day of some hypothetical Google rollout.

Second reason: precisely because the measurement is still rare, it is a window of advantage. Most companies do not know whether AI engines cite them, because they have never looked. Those who audit, then correct, and build their authority now will be hard to dislodge once the topic becomes obvious to everyone. A GEO audit is, first of all, giving yourself the eyes to see a field your competitors are still crossing blind.

What a GEO audit measures: the 3 dimensions

A GEO audit scores three cumulative dimensions: findability (are your pages indexed and accessible to AI engine crawlers), extractability (does your content offer self-contained passages that directly answer a question), and credibility (are your claims backed by named sources and verifiable data). Failing a single one of these dimensions is enough to not be cited.

These three dimensions are not an arbitrary split: they follow the way an assistant builds its answer. It first retrieves candidate pages through a web search, reads their content, extracts passages from them, then cites what it actually used. Each step has its own filter, and therefore its own axis of audit.

1. Findability
   Your pages have to be retrievable by the engines the AIs rely on. ChatGPT, for example, draws on the Bing index complemented by its own OAI-SearchBot crawler. The audit checks indexation, the absence of blocks on the useful crawlers in the robots.txt file, and the presence of content in the rendered HTML rather than hidden behind unexecuted JavaScript. It is the prerequisite nothing makes up for: a page invisible to retrieval will never be a candidate. To stay on the safe side, run a quick indexation check on your key pages before anything else.
2. Extractability
   The model does not lift a whole page, it draws out passages. The audit therefore assesses whether your content offers self-contained paragraphs that open with a direct answer, understandable without reading what came before. This is the most actionable lever: an answer diluted into the third paragraph is bad for extraction, while an answer placed at the top of a section is citable.
3. Credibility
   At equal relevance, the model favours content backed by named sources and quantified data. The audit records the share of your claims that stay unproven. This is the point the foundational academic study measured as one of the most decisive: adding quantified citations and credible sources can lift visibility in generated answers in a measurable way, where the keyword stuffing inherited from old SEO has no effect. This is exactly what our guide on E-E-A-T and AI content extends.

SEO audit and GEO audit: what changes

An SEO audit assesses your ability to rank well in Google's list of links; a GEO audit assesses your ability to be cited in an answer written by an AI. Both share common foundations, but a GEO audit gives more weight to the extractable passage and the named source than to position alone. They complement each other: you run GEO on top of SEO, not instead of it.

The costliest confusion is believing that "being first on Google" amounts to "being cited by ChatGPT". The two goals overlap on the foundations, reliable content, clean technical access, authority, but they diverge on what carries the final decision.

The most structuring row is the second to last: the measurement. Google provides Search Console and its numbers. AI engines offer no mainstream equivalent. The only reliable way to know whether you are cited is to ask the question yourself and look. It is a real constraint, but it is also why a serious GEO audit has value: it does, methodically, what few take the time to do. To position the two disciplines, see our GEO vs SEO comparison and our analysis of how AI answers cut traffic to classic sites, with the data to back it.

How a GEO audit unfolds

A GEO audit unfolds in four steps: we build the list of your real business queries, put them to the AI assistants with web search switched on to record your starting citation rate, analyse the causes across the three dimensions, then rank the corrective actions. This is the method Cicero Studio applies: GEO audit, editorial production, automated semantic internal linking.

1. Frame the business queries

We draw up a list of ten to twenty questions your customers genuinely ask before they buy or choose. Not generic keywords, but questions phrased the way a human puts them to an assistant. This is the step that decides the relevance of everything else: auditing the wrong queries gives a worthless result.

2. Record the starting citation rate

We put each query to the assistants with web search switched on, and we note query by query: is your brand cited, who else is, and in what form (link, mention, paraphrase). This record gives an honest photograph of your starting point, far more useful than a composite index. Because the same question can yield different answers from one session to the next, we test cleanly before recording the variations.

A GEO audit is discussed as much as it is measured: business queries come from the people who talk to customers.

3. Analyse the causes across the three dimensions

For each query where you are absent, we trace the cause: were you simply invisible to retrieval (findability), present but not extractable (structure), or extractable but less credible than a sourced competitor (credibility). It is this cause-by-cause diagnostic that turns an observation into an action plan.

4. Rank the actions

We sort the fixes from the most profitable to the heaviest: first the technical blocks that exclude you outright, then the quick editorial restructurings, finally the deeper work such as building a coherent set of content on your topic. We always start with what unlocks the most citations for the least effort.

How to read the results

You read a GEO audit by first looking at the starting citation rate, query by query, then the cause of each absence across the three dimensions, and finally the priority of the actions. A good deliverable answers three questions: where am I absent? for what reason? and what will unlock the most citations first?

The trap, when reading, is to fixate on a single global figure. The average citation rate is useful as a benchmark, but what matters is its breakdown. Being absent on queries where your competitors are cited is a strong signal; being absent where no one is cited is far weaker. The same nuance applies to the causes: a technical block is fixed in hours, topic authority is built over months.

Read properly, then, a GEO audit is not a report card: it is a map. It tells you where the gaps are, which ones cost the most, and where to start to close them. The rest is execution.

What to do after the audit

After the audit, you execute in priority order: lift the technical blocks that exclude you from retrieval, restructure existing content into direct answers by question, anchor each claim in a named source, then build a connected set of content that establishes your topic authority. You re-test the queries at regular intervals to track progress.

The audit is only worth what follows it. A technical block lifted can bring a page into the day's candidate pool overnight. A direct answer added at the top of a section can make it extractable in an hour of work. But durable presence in AI answers does not come from an isolated page: it comes from a network of content that reinforces itself and signals your expertise on a topic, a pillar that sets the theme and satellites that dig into it, linked by contextual internal links.

This is exactly how Cicero Studio works: a GEO audit that measures your current citability, a human editorial production assisted by AI that creates the missing content, and an automated semantic internal linking that makes it all work together. Agency-quality work, software-grade productivity. To go further on execution, see our practical method for appearing in ChatGPT and Google AI Overviews, our GEO guide for small businesses and our overview of what a GEO agency actually does.

Alexis Dollé

CEO & Founder of Cicero Studio

I audit AI visibility by testing it by hand, query after query, across dozens of sites on their real business questions. This page sums up what I see day to day at Cicero Studio. My conviction: a GEO audit only makes sense if it leads to prioritised actions and a repeated measurement, not to a score that merely reassures.

LinkedIn →

What this page does not cover

For the sake of honesty, and because it is exactly the kind of transparency AI engines reward, here are the limits to know before you build a strategy around a GEO audit. Stating what a method does not do is often better than overselling what it does.

Going further

We document our approach in the open, because it is our best proof. Each resource below digs into a precise angle of the audit and of visibility in AI engines, from the quantified method to the mechanics of citation, by way of the small-business case. Here are the most useful pieces depending on what concerns you: