AEO Audit: A Practical AI Visibility Checklist

An AEO audit shows what answer engines say about a brand, what they get wrong, and which competitors they recommend instead. It is not a checklist of keywords. You need a repeatable set of buyer questions, responses captured in comparable conditions, and a way to separate a genuine mention from an answer that only repeats the brand name you supplied.

In short: what an AEO audit checks

An AEO audit tests how answer engines respond to relevant questions about a category, a problem, and a brand. It records spontaneous mentions, recommendation position, competing brands, factual errors, cited URLs, and the difference between a model's native answer and its web-enabled answer. In our 2026 sample of 104 audits and 6,820 AI responses, only 16% of brands appeared spontaneously. That is a measured sample, not a market-wide benchmark.

What is an AEO audit?

An AEO audit is a structured review of a brand's presence in answer-engine responses. It answers a practical question: when a prospective buyer asks an AI for options, advice, or a comparison, does the answer identify the brand accurately and in the right context?

It differs from a conventional search audit because the unit of analysis is the answer, not only a ranking URL. A useful audit looks at the wording of the question, the full response, the sources named by the engine when available, and the brands that appear beside yours. For a narrower definition, see our AEO audit glossary entry.

When an AEO website audit is worth doing

Start when AI-assisted research matters to the people you want to reach, or when clients report unfamiliar claims and recommendations in AI answers. The goal is not to chase every model or every prompt. It is to establish a baseline for the questions that influence a real decision.

• Before a content or digital PR project: document the starting point, including missing or inaccurate claims.
• After a site or positioning change: check whether the engine still describes the offer, category, and evidence correctly.
• For an agency client: compare the brand against the competitors repeatedly recommended for the same use case.
• When a factual error is found: preserve the prompt, model, mode, response, and cited source before proposing a correction.

AEO audit methodology: six steps

1. Define the decision and the audience

Write down the decision the answer should support: selecting a provider, shortlisting tools, finding a local service, or checking a claim. Then define the market, language, audience, and category. A prompt such as "best accounting software for a UK freelancer" is not interchangeable with "what is accounting software?"

2. Build a small, defensible prompt set

Use three prompt types: unbranded category questions, comparison questions, and brand-verification questions. Keep the wording natural and save the exact text. A branded prompt can test accuracy, but it cannot prove spontaneous visibility because it has already introduced the brand.

3. Test native and web-enabled answers separately

Where the engine offers both modes, record them separately. The native answer reflects the model response without a live search step; the web-enabled answer can introduce current pages and citations. Combining the two hides a useful diagnostic: a brand may be known by a model but absent from current web results, or the reverse.

4. Capture the evidence, not only a score

Save the prompt, model, mode, date, complete answer, named competitors, cited URLs, and any uncertainty stated by the engine. This makes the audit reproducible. It also stops a flattering but invented claim from being treated as a win.

5. Measure the signals that explain the result

Signal	What to record	Why it matters
Spontaneous mention	Whether the brand appears in an unbranded question	Separates category recognition from name repetition
Recommendation position	Where the brand appears in a shortlist or comparison	Shows how the answer frames the brand against alternatives
Answer accuracy	Claims that match, omit, or contradict verified information	Turns vague reputation concerns into reviewable evidence
Competitive presence	Which alternatives recur across the same question set	Reveals the shortlist the engine is actually constructing
Source and URL quality	Cited pages, unavailable pages, and hallucinated URLs	Points to the content or entity evidence that needs work

6. Turn findings into a bounded action list

Link each action to the evidence that produced it. A missing product fact calls for a clear, accessible source page. A recurring wrong claim calls for a correction path and better corroborating sources. A competitor repeatedly cited for a use case may expose a content, entity, or authority gap. Do not promise that any one edit will change a model response.

What an AEO audit report should contain

A report should be usable by the person who owns the website, the brand, and the content roadmap. Include the prompt set, testing conditions, a response-level evidence log, a competitor view, factual-risk examples, and priorities tied to expected business decisions. If an audit tool produces a single number without the underlying responses, it is hard to challenge or act on.

AI Labs Audit is designed to retain that evidence across models and to distinguish spontaneous visibility from answers that merely echo a supplied brand name. It also flags invented URLs and claims so a positive-sounding answer is not automatically scored as accurate.

Common AEO audit mistakes

• Counting every branded mention as visibility.
• Changing the prompt wording between models, then comparing the results as if the test were identical.
• Mixing a model's native answers with web-enabled answers in one metric.
• Recording only the first sentence and missing a correction, caveat, or competitor later in the answer.
• Treating an AI-generated claim as proof without checking it against first-party information.

AEO audit checklist

• Define the audience, market, language, and decision behind each question.
• Use unbranded, comparison, and brand-verification prompts.
• Record the model, mode, date, and full answer for every test.
• Separate spontaneous mentions from prompted name repetition.
• Check recommendation position, competitors, factual accuracy, and cited URLs.
• Prioritize actions that have a named evidence trail and an owner.

Frequently asked questions about AEO audits

What is the difference between an AEO audit and an SEO audit?

An SEO audit reviews how search engines discover, index, and rank pages. An AEO audit reviews the answers generated from a question, including mentions, recommendations, claims, competitors, and sources. The two can inform each other, but they are not interchangeable.

Can an AEO audit prove that a brand will be recommended?

No. AI responses vary by model, mode, query wording, time, and context. A sound audit documents the conditions of the test and looks for repeated patterns rather than making a guarantee from one response.

Which answer engines should an AEO audit include?

Choose engines that match the audience and market you are investigating. The important point is consistency: use the same prompt set and capture the testing conditions for each engine included.