Skip to content
Writing
getting started

Check whether AI names you. Free, three ways.

You can do this by hand in an afternoon, with a spreadsheet in a week, or in under ten minutes with no account. All three are listed here, including the two that do not involve us.

Published 2026-08-19 | 4 minute read

The question is simple: when somebody who needs what you sell asks an assistant, does your name come up? Finding out is free. Finding out in a way you can trust takes a little care, and the care is almost entirely about avoiding four mistakes.

The four mistakes

  • Asking about your brand by name. The model will find you. Your customers do not search that way, so the answer tells you nothing about whether you get recommended.
  • Asking from your own account. History and location bend the reply toward you. You are reading a version no prospective customer will ever see.
  • Asking once. Repeat asks of the same question return different names. One reading is an anecdote.
  • Asking one assistant. They disagree with each other far more than they disagree with themselves, and your buyers are not all in the same one.

Option one: by hand, about an hour

Write five questions the way a buyer would type them — the problem, not your category jargon, and not your name. Open each assistant in a private window, signed out, with the location set to your market. Ask all five in each. Paste every answer into a document with the date and the engine beside it.

Then count two things: how many of the answers named you, and which other companies appeared. The second list is more useful than the first. It is your actual competitive set as an assistant sees it, and it usually contains at least one name you were not tracking.

Keep the raw text, not your summary of it. The paragraph is the evidence; a tally you wrote from memory is not something you can re-check in a month.

Option two: a spreadsheet you keep

The hour above tells you where you stand today. It does not tell you whether you moved, and that needs the same questions asked the same way on a schedule. One row per question per engine per date, with the answer text in a cell and a column for who was named. Repeat monthly.

This works and it is genuinely free. What it costs is discipline, and the failure mode is specific: the question set drifts. Somebody rewords a prompt, and this month stops being comparable with last month without anybody noticing. Freeze the wording, and change it only by starting a new sheet.

Option three: the free scan

Five real questions, sent live to an engine with web search on, with the whole answer text kept. It takes under ten minutes, needs no account and no card, and it exists because the hour above is a real hour and most people do not have it.

It also does the two things a private window cannot: it asks from a clean context every time rather than from a browser you have used, and it stores the result with its date and model so the next reading is a comparison rather than a memory.

Whichever way you do it, one reading establishes a position and not a trend. Nothing honest can report a change until the same questions have been asked at least three times.
questions

The short answers

How can I check my AI visibility for free?

Three ways. By hand in about an hour: five buyer-style questions, asked signed out in a private window across each assistant, with the answers saved. With a spreadsheet you keep monthly, which is the same thing on a schedule. Or with a free scan, which takes under ten minutes and needs no account.

Should I ask ChatGPT about my brand by name?

No. It will find you, and the answer tells you nothing. Ask the question a customer would type — the problem they have, not your company or your category jargon — because that is the query where being named actually matters.

Why should I sign out before checking?

Your account carries chat history and a location, and both bend the reply toward you. Checking from your own logged-in session is the single most common way people conclude they are visible when a stranger asking the same thing gets two other names.

How many questions do I need to ask?

Five is enough to stop being an anecdote and few enough to actually do. What matters more than the count is that they are different phrasings of a real buying question and that the wording stays frozen between readings.

Do I need to check more than one AI assistant?

Yes. Assistants disagree with each other considerably more than they disagree with themselves, so a reading from one of them is a reading about one of them. Your buyers are spread across several.

Is a free scan enough, or do I need a paid tool?

A free scan answers "where do I stand". It cannot answer "did I move", because that needs the same questions asked repeatedly with a variance test on top. If the answer to the first question is uncomfortable, the second one becomes worth paying for; if it is not, it does not.

Find out in under ten minutes

One free check, no account and no card. You get the answers the assistants gave, not a score.

Every figure quoted here comes from a real question put to a real assistant, saved with the date and the assistant that answered it. This piece was written to answer the search "check ai visibility free".

free scan

are you in the answer?

Five real questions against a live answer engine with web search on, each asked ten times. That is fifty real calls, so the report lands in under ten minutes rather than while you wait. You see every question before anything runs, and you get the answer text, who got named and the sources the engine read, not a score.

your site/what you sell/where your buyers are/the questions