Matt Turley, Continuum. Updated October 2, 2026.
What a good audit checks, which questions it should ask, why one run is not enough, what it should cost, and what the report looks like.
An AI visibility audit checks whether ChatGPT, Perplexity and Gemini name your business when your customers ask them who to hire, who they name instead, and which websites they read before answering. A good one asks 10 to 15 questions in your customers' words, asks each one three times on each tool, and hands you the three fixes worth the most. Ours is the AI Answer Check: $99, delivered within 3 business days or it is free.
Three things, and the first one is the headline.
A fourth line item is worth having: any answer that states a wrong fact about you, like a price you do not charge or a service you do not offer. Cited is not the same as named, and named is not the same as described correctly. More on the difference between the first two is in how to rank in ChatGPT for your business.
Your customers' words, not keywords. Nobody types “plumbing services Dallas” into ChatGPT. They type a sentence with a problem in it. Five to ten questions is enough for a first read, 12 to 15 is what our report uses, and none of them should contain your business name, or the tool will simply look you up.
The questions are frozen before the first run and never edited afterwards, so the re-measure a month later is like for like. For a Dallas plumber, three from the set we published look like this:
The whole list of 10, and what the AI tools answered, is on the plumbers in Dallas research page.
Because answers move between runs, and the tools read different sites. Ask the same question twice and you can get two different lists. One screenshot proves nothing either way.
The tools disagree more than people expect. In our October 2026 Dallas research, across 12 trades and 720 answers, 691 businesses were named on at least one question check by Perplexity or Gemini. Only 170 of those 691 were named by both tools. In a September realtor check, ChatGPT cited one rankings site, therealtorrankings.com, in 11 of 12 answers, while Gemini leaned on a different site entirely, and only 1 of ChatGPT's 70 citations was an agent's own website. Measure one tool and you know about one tool.
Three runs give you a majority rule: named in 2 of 3 counts, named in 1 of 3 does not. Three tools give you the whole picture a customer might see.
One vendor's published benchmark, The Answer Engine (June 21, 2026, linked in the sources), puts a one-time AEO audit at $800 to $2,500, a starter retainer at $2,500 a month, and calls $500 to $1,000 a month a “below-market danger zone”. That is their view and they are entitled to it. Here is ours, with the report in view so you can judge for yourself.
| What | Price | What you get |
|---|---|---|
| AI Answer Check (the audit) | $99 | Your buyers' questions on ChatGPT, Perplexity and Gemini, 3 runs each: named or not, who is named instead, which sites they read, wrong facts stated about you, and the three fixes worth the most. Read by Matt before it goes out. In your inbox within 3 business days of your completed order or it is free; full refund within 7 days if it was not useful. |
| Get Users, Proving rate | $500/mo, month to month | Each week, done for you once you approve: page rewrites titled to the exact questions, the directories and best-of lists to get onto, pitched and followed up, Reddit replies drafted for you to post, and a weekly scoreboard of what the three tools say about you. Prefer to keep the keys? Same price, you make the changes. |
| Get Users, Founding Client rate | $999/mo, locked for life | The same work once your written goal is hit on two weekly reads in a row. Standard rate for new clients is $1,999/mo. |
Why ours is cheaper than that benchmark: the probing is automated. The same system that runs your $99 report runs every week for Get Users clients, so the cost of collecting every answer and logging every citation is small. The part that is not automated, reading the answers and choosing the three fixes, is done by one person, me, and I read every report before it is sent. The exact terms are on the guarantees page, and every price on this site comes from one pricing page.
One table per AI tool, one row per question. For each of your questions it shows whether ChatGPT, Perplexity and Gemini named you in most of their three runs, whether they linked your site as a source, and which businesses they named instead. Then the sites the answers cited most, any wrong facts the tools state about you, and the three fixes worth the most.
Named and linked are kept apart on purpose. A tool can use your site as a source and still recommend someone else, and that gap is usually where the first fix is. What is in the report, question by question, is on the product page.
Three fixes, ranked by how much of the citation share they touch. The sites the answers read are the lever, and which kind of site matters depends on what you sell. In the same Dallas research, directories and review sites were 30.7% of citations for the 12 local trades, while best-of lists were only 10.8%. For AI chatbot tools it flipped: best-of lists and comparisons were 54.3% of citations and directories 2.6%.
Yes. Write down ten questions in your customers' words, without your name in them. Tonight, ask each one three times in ChatGPT with web search on, then in Perplexity, then in Gemini. For each answer write down whether you were named, who was, and every site it cited. Only count a question as won when you are named in 2 of 3 runs.
The caveat is the volume: 10 questions times 3 tools times 3 runs is 90 answers to read and log by hand, and the sites cited need sorting into types before the pattern shows. Our free AI visibility check runs a few of your questions live and scores how often you show up, which is a fair first look. The $99 report is the full version with the competitors, the sites and the fixes.
The AI Answer Check is the audit described on this page: your buyers' questions, three AI tools, three runs each, who is named instead of you, which sites they read, and the three fixes worth the most. It is $99, ordered online in about two minutes, and it does not promise what any AI tool will say.
Get Users is the monthly version: the Proving rate is $500/mo while we prove the method works for you, then the Founding Client rate, $999/mo locked for life once the written goal is hit (standard $1,999/mo), month to month, with a weekly scoreboard.