Short answer
An AI visibility audit measures whether AI assistants — ChatGPT, Perplexity, and Google's AI Overviews — name your company when buyers ask them a question, how often you appear versus competitors, and whether what they say about you is accurate. It is not a rankings report; it inspects the answer itself. A good audit checks brand mentions across a panel of realistic prompts, the sources AI cites, NAP consistency, structured data, third-party mentions, and review signals. You can run a 30-minute version yourself with free tools, or hire an agency for scale and fixes. LYVIA's 2026 Barometer found 37.7% of 390 French SMBs were invisible to AI assistants. The audit's value is the fix list it produces, not the score.
What an AI visibility audit is
An AI visibility audit measures whether AI assistants like ChatGPT, Perplexity, and Google's AI Overviews mention your company when a buyer asks them a question — and whether what they say is accurate. It is not a website crawl or a rankings report. It is a snapshot of how large language models describe your brand, which competitors they name instead of you, and which sources they pull that description from.
Traditional SEO asks where you rank on a page of blue links. An AI visibility audit asks a different question: when the answer is generated rather than listed, are you inside it? The audit captures three things — presence (are you mentioned at all), share of voice (how often you appear versus rivals), and accuracy (whether the AI's version of your services, location, and pricing is correct).
What an audit actually measures
A complete audit checks the concrete signals that AI engines use to build an answer about your company. Each check maps to something you can fix.
- Brand mentions: whether your name surfaces across a panel of realistic buyer prompts, not just one.
- Citation sources: which pages the engine links or footnotes — your own site, directories, review sites, or a competitor's blog.
- NAP consistency: whether your name, address, and phone match across the web, since conflicting data makes AI hedge or omit you.
- Structured data: whether your pages expose schema that machines can read cleanly.
- Third-party mentions: independent articles, listings, and profiles that AI treats as corroboration.
- Review signals: volume, recency, and rating, which shape how confidently AI recommends you.
Read together, these signals explain not just whether you are visible, but why.
How to run a 30-minute audit yourself
You can run a useful AI visibility audit yourself in about 30 minutes with nothing but the free tiers of ChatGPT, Perplexity, and Google. The goal is a repeatable baseline, not a single lucky answer.
The 30-minute method: build a panel of 8 to 12 buyer prompts — the real questions a customer would type, such as "best [your service] in [your city]" or "who should I hire for [problem]". Run every prompt in all three engines. Record, in a simple table, whether you appear, who appears instead, and which sources are cited.
Look for patterns, not single results. If you are absent from most prompts in two of three engines, you have a presence problem. If you appear but the details are wrong, you have an accuracy problem. If a competitor is cited from a source you could also earn — a directory, a roundup, a review site — you have a clear, fixable gap. For the full step-by-step version, run the complete audit method.
What LYVIA's 2026 Barometer found
LYVIA's 2026 Barometer found that 37.7% of the 390 French small and mid-sized businesses it studied were effectively invisible to AI assistants — never surfaced in a relevant answer. The full methodology is published in the study.
We cite this as our own research, not an industry benchmark: it reflects a specific sample of French SMBs at a specific moment. Still, the direction matters for US and UK owners too. If more than a third of a studied market simply does not appear when buyers ask AI for a recommendation, the cost of skipping an audit is measured in conversations you never knew you lost.
What agencies charge for an audit
What agencies charge for an AI visibility audit varies widely, and any firm quoting a single fixed number is guessing. Based on proposals we have seen and engagements we run, three models are common.
- One-off audit: a fixed-scope report with prompt panels, findings, and a fix list. In our experience these typically range from roughly a few hundred to a few thousand dollars, depending on the number of engines, prompts, and competitors tracked.
- Retainer: a monthly fee that re-runs the audit and tracks movement over time, since AI answers shift. Priced by cadence and scope.
- Included in a GEO engagement: the audit is bundled as the diagnostic step of a broader visibility program, so you pay for the fixes, not the report.
Treat any figure — ours included — as a starting range, not a quote. Scope drives price far more than the name on the invoice does.
When to hire an agency versus DIY
Hire an agency when the stakes or the scope outgrow a 30-minute self-check. Do it yourself when you are still building a baseline and have someone with a few hours to spare.
Clear triggers to buy rather than build:
- AI answers actively state something wrong about your company, and it is costing you leads.
- You compete across several cities or service lines and need a structured panel, not a handful of prompts.
- No one on your team has time to run the audit monthly and act on it.
- You want the fixes delivered, not just the diagnosis — schema, content, citations, and reviews handled end to end.
LYVIA is a Paris-based agency serving international clients, so a US or UK owner can get an outside read without adding headcount. The build-versus-buy line is simple: self-run to learn, hire to move the numbers.
What an audit does not tell you
An AI visibility audit is a snapshot, not a guarantee. It tells you how engines describe you today, in this sample of prompts — and both of those change.
Three limits to keep in mind. AI answers are non-deterministic: the same prompt can return different results minutes apart, so one answer is never proof. Engines update their models and sources without notice, so a clean audit can drift within weeks. And an audit measures visibility, not revenue — being cited is necessary, but it does not by itself close the sale. Read the audit as a compass, not a scoreboard.
Common mistakes when measuring AI visibility
Most flawed AI visibility audits fail for the same handful of reasons. Avoiding them is what separates a real baseline from a false comfort.
- One prompt: a single query tests luck, not visibility. Use a panel.
- One engine: ChatGPT, Perplexity, and AI Overviews draw on different sources and disagree often.
- No baseline: without a dated record, you cannot tell whether a fix worked.
- Treating one answer as truth: non-determinism means you need repeats, not a screenshot.
- Ignoring sources: the citation list tells you where to earn your way in — skip it and you are guessing.
How to turn findings into a plan
Turn audit findings into a plan by mapping each gap to a specific fix, then working in order of leverage. The audit's value is the to-do list it produces, not the score.
- Wrong or missing facts: correct your structured data and key pages so machines read you cleanly.
- Absent from answers: earn the sources AI already cites — directories, roundups, and review sites.
- Beaten by a competitor: build content that answers the exact buyer prompts you lost.
- Thin trust signals: raise review velocity and third-party mentions.
To go deeper on being quoted, see how to get cited by ChatGPT and how to appear in Google AI Overviews.
Frequently asked questions
What is an AI visibility audit?
An AI visibility audit is a structured check of whether AI assistants — ChatGPT, Perplexity, and Google's AI Overviews — mention your company when buyers ask them a question, how often you appear versus competitors, and whether what they say about you is accurate. Unlike a traditional rankings report, it inspects the generated answer itself rather than a page of links. It reviews brand mentions, cited sources, NAP consistency, structured data, third-party mentions, and review signals, then turns those findings into a prioritized list of fixes you can act on.
How much does an AI visibility audit cost?
It depends heavily on scope, and any firm quoting one fixed number is guessing. Based on proposals we have seen and engagements we run, a one-off audit typically ranges from roughly a few hundred to a few thousand dollars, driven by how many engines, prompts, and competitors you track. Retainers that re-run the audit monthly are priced by cadence. Some agencies bundle the audit into a broader visibility program, so you effectively pay for the fixes and get the diagnostic for free. Treat every figure as a starting range, not a firm quote.
Can I audit my own AI visibility for free?
Yes. You can run a solid baseline in about 30 minutes using only the free tiers of ChatGPT, Perplexity, and Google. Build a panel of 8 to 12 realistic buyer prompts, run each one in all three engines, and record whether you appear, who appears instead, and which sources are cited. Look for patterns across the panel rather than trusting one answer, since results vary. A free self-audit will not scale to many cities or service lines, but it is more than enough to tell whether you have a presence, accuracy, or sourcing problem.
What tools are used for an AI visibility audit?
The core tools are the AI assistants themselves — ChatGPT, Perplexity, and Google's AI Overviews — because you are testing what they actually say. Beyond that, you need a simple table or spreadsheet to log prompts, appearances, competitors, and cited sources so you have a dated baseline to compare against later. Agencies layer on prompt-tracking dashboards, schema validators, and review-monitoring tools to run larger panels and track movement over time. For a self-run audit, though, the free engines plus a spreadsheet cover the essentials.
How often should I run an AI visibility audit?
Run a light audit monthly and a fuller one each quarter. AI answers are non-deterministic and the engines update their models and sources frequently, so a result you captured last month can drift without any change on your side. A monthly spot-check on your core prompts catches regressions early, while a quarterly deep run — more prompts, more engines, competitor tracking — measures whether your fixes are moving the numbers. Always keep a dated record; without a baseline you cannot tell improvement from noise.
How is it different from a traditional SEO audit?
A traditional SEO audit measures where your pages rank in a list of links and inspects technical health, keywords, and backlinks. An AI visibility audit measures whether you appear inside a generated answer — what ChatGPT or Perplexity actually tells a buyer — and whether that description is accurate. The two overlap on signals like structured data, citations, and reviews, but the questions differ: SEO asks about position on a results page, while an AI visibility audit asks whether the model names you at all and gets your details right when it does.
Want an outside read on how AI assistants describe your brand? LYVIA is a Paris-based agency running AI visibility audits for international clients, and we turn the findings into fixes. Book a call
