AI brand monitoring: how to track your brand’s mentions in AI answers

AI brand monitoring means checking, on a schedule, whether AI tools like ChatGPT and Google’s AI Overviews name your brand when buyers ask. You can do it by hand with a spreadsheet or with a tool, but ask each question more than once. In our study, the brands AI Overviews named overlapped only 60.06% across three runs of the same search.

What is AI brand monitoring?

It’s like keeping an eye on your reviews or press coverage, but for AI answers. Buyers now ask AI what to buy and who to hire. AI brand monitoring tells you what those answers say about you.

For each question you track, you want to know six things:

• Named or not. Does the answer mention your brand at all?
• Position. Are you the first pick, or the last name in a list?
• What it says. Are the facts right? Is the tone good, bad or neutral?
• Competitors. Which other brands are named instead of you, or next to you?
• Sources. Which pages does the AI link to? These are the sources it used.
• Your site. Does the answer link to your own website?

Why you can’t see this in your normal reports

Your analytics show visits, not mentions. And many AI mentions never lead to a visit.

• In our study, when ChatGPT named brands it already knew, 85% of those mentions had no link to the brand’s own website. No link means no visit, so the mention never shows up in your analytics.
• When Google’s AI Overviews named a brand, 80.69% of the links next to it went to other websites, not the brand’s own site.
• Google Search Console counts clicks from AI Overviews and AI Mode with your normal web search data. It doesn’t tell you whether the AI named your brand.

So the only way to know what AI says about you is to read the answers.

Why one check is not enough

AI answers change, even when you ask the exact same thing. We tested this in our study of 10,000 AI answers. We searched 1,000 keywords in Google three times each:

In our study, an AI Overview showed on only one or two of the three searches for 7.9% of keywords, and on none of them for 9.4%.

The pages Google used hardly changed. The brands it named changed a lot more. For one search about tattoo lotion, the first answer named two brands. The next two answers kept one of them and swapped the other for a different brand.

So a single check can mislead you. If you check once and see your brand, you may only be named some of the time. If you check once and don’t see it, you may still show up on other runs. Ask each question at least three times, and track how often you’re named.

Answers also differ by platform and by person. In our study, ChatGPT and AI Overviews shared only 3.33% of the pages they linked to, so track each AI on its own. ChatGPT can also use a person’s rough location and, if memory is on, what it remembers about them. The answer you see may not match the one your buyer sees.

How to track brand mentions in AI by hand

Before you build your own sheet, you can get a first snapshot from our free AI visibility audit. It shows whether AI recommends you or your competitors, and what your buyers ask AI. To track it yourself, follow these steps:

1. Write 20 to 30 buyer questions. Use your buyers’ own words. Mix questions from each step of buying, like “best [product] for [need]”, “[brand] vs [brand]”, “who does [service] near me?” and “is [your brand] any good?”.
2. Pick the AIs your buyers use. Start with ChatGPT and Google’s AI Overviews. Add AI Mode, Gemini or Claude if your buyers use them.
3. Use a clean session. Log out, or use a private window, so your own history doesn’t shape the answer. Add a city to the question if location matters to your business.
4. Ask each question three times. In ChatGPT, use a new chat each time. In Google, search again in a fresh private window.
5. Log every answer in a sheet like the one below.
6. Repeat every month with the same questions, so you can compare.

The right questions depend on your business. We cover this for AI SEO for contractors, AI SEO for ecommerce and AI SEO for SaaS.

Here is a simple sheet you can copy:

How to track brand mentions in ChatGPT

Open a new chat for each run. For many questions, ChatGPT searches the web on its own. When the answer has a Sources button, open it and write down the sites. If ChatGPT answers without searching, it’s relying on what it learned in training. Note that too. It tells you whether ChatGPT knows your brand without looking it up.

In your analytics, look for visits tagged utm_source=chatgpt.com. ChatGPT adds this tag to the links it shows, so you can count the visits it sends you. Just remember that many mentions never lead to a click.

How to track AI Overviews and AI Mode

Search in a private window. First, note whether an AI Overview shows at all. If it does, open the full summary and write down the brands it names and the pages it links to. For AI Mode, ask the same question there and log it in its own rows. The two can give different answers.

How to track your brand with AI monitoring tools

Checking by hand works for a short list. For more questions, more AIs or weekly checks, a tool saves time. AI monitoring tools send your list of questions, often called prompts, to several AIs on a schedule. They save the answers and show your mentions, position, competitors and linked pages over time.

What to look for in a tool:

• The AIs your buyers use. Make sure it covers ChatGPT, Google’s AI Overviews and AI Mode, plus any others you care about.
• Repeat runs. It should ask each question more than once, or often enough to show a trend.
• Location. It should let you pick the country or city your buyers are in.
• Sources. It should show the pages the AI linked to, not just whether you were named.
• Your own questions. You should be able to add the exact questions your buyers ask.
• Export. You should be able to download the data.

Know the limits, too. A tool only tracks the questions you give it, and real buyers word things in many ways. Its answers are samples, not a full count of what every buyer sees. We compare the options in the best AI visibility tools.

AI share of voice: how to measure it

AI share of voice is your share of all the brand mentions in AI answers for your list of questions. It shows how you compare with competitors, not just whether you’re named.

To work it out:

1. Count every brand mention across all your answers. Count each brand once per answer.
2. Count how many of those mentions are your brand.
3. Divide your mentions by the total, then multiply by 100.

For example, say your answers name brands 200 times in total, and 30 of those are your brand. Your share of voice is 30 ÷ 200 × 100 = 15%. Track this number every month, for each AI, and next to your main competitors.

What to do with what you find

Your tracking sheet shows you where to act:

• You’re not named at all. Look at the pages the AI linked to. Those are the pages to get your brand onto. Start with how to get recommended by ChatGPT and how to show up in Google’s AI Overviews.
• You’re named, but the facts are wrong. Fix the facts on your own site first. Then check the pages the AI linked to for old or wrong details.
• Competitors are named more often. Compare the pages that name them with the pages that name you. The gap is your to-do list.
• You’re named on some runs but not others. You’re close. More mentions on the pages the AI uses give you more chances to be picked.

Get your free AI visibility audit

Not sure what AI says about your brand today? Our free AI visibility audit shows where you stand in AI, what your buyers ask AI, and how to close the gap.