ChatGPT visibility tracker: how to track brand mentions in ChatGPT
ChatGPT has no results page, so a ChatGPT rank tracker can't work the way a Google one does. What you can measure is how often ChatGPT names your brand across a fixed set of buyer prompts, what it says about you and which pages it cites.
ChatGPT rank tracking vs visibility tracking
Searching for a "ChatGPT rank tracker" suggests a stable position you can follow week by week, and the evidence says that position doesn't exist. In SparkToro's research with Gumshoe, each prompt ran 60 to 100 times per platform. The odds of two answers sharing the same list of brands were under 1 in 100, and the same list in the same order came up about once in 1,000 runs.
The authors' conclusion is blunt: any tool promising a "ranking position in AI" is "full of baloney", while visibility percentage across dozens to hundreds of prompts, each run several times, is a reasonable metric.
Average position still has a use. Across many answers it tells you whether ChatGPT tends to name you early or late, while on one prompt in one run it tells you nothing. Lead your reports with visibility and treat position as supporting detail.
What a ChatGPT visibility tracker records
For each prompt and run, a tracker should store the full answer and then pull these out of it:
- whether your brand appeared, counting the other spellings people use
- every other brand named, including ones you never thought to track
- the tone of the sentence about you
- all cited URLs, since ChatGPT draws on web search for most commercial questions
- the date, country and collection method, so that two runs can be compared fairly
The full answer matters more than the labels. Automated sentiment gets things wrong, and when a chart says your tone dropped, you need the sentence that caused it.
Prompt set built from SEO keywords
Keywords are a good raw material and poor prompts. Nobody types "payroll software" into ChatGPT and stops there. They describe their situation, so convert each keyword into the question behind it:
| SEO keyword | Prompt to track |
|---|---|
| payroll software | I run a 12-person design studio in Texas. Which payroll software should I use? |
| gusto alternatives | Gusto is getting expensive for us. What should a small agency switch to? |
| payroll compliance | What's the easiest way to stay compliant with payroll tax in two US states? |
Start with 20 to 30 of these and keep your brand name out of them. A prompt that contains your name proves nothing about discovery. Keep a few branded prompts ("is X good for agencies?") in a separate group to check how you're described.
Leave the set alone once it's running. Adding prompts changes your baseline, so your chart ends up measuring your edits. AI Monitor's prompt generator does the keyword conversion for you, from your site or a competitor's, and you can edit the results before a report runs.
See your own numbers. Create a free account, paste ten prompts and get a first ChatGPT report in minutes. New accounts get free credit and no card is needed.
Tracking ChatGPT mentions by hand
You can do this today with a temporary chat, which keeps memory and past conversations out of the answer. Ask each prompt, paste the answer into a spreadsheet, then mark mentions and cited domains. Do three rounds on different days, for the reasons SparkToro's numbers make obvious.
Twenty prompts take about an hour per round, once you add the logging. Expect to give up around the time you add a second country or a third assistant. Our AI visibility tracking guide covers the spreadsheet and the formulas in more detail.
Logged-out sessions, memory and personalization
Every tracker we know of measures a logged-out or clean session, and real buyers usually aren't in one. ChatGPT users have memory, custom instructions and a chat history that can shift the answer toward brands they've already discussed.
No outside tool can see that, and we don't expect that to change, since doing so would mean reading private accounts. A clean-session result is a benchmark of what the product says to someone it knows nothing about. It's the fairest comparison between brands that we have. It still isn't the answer your customer saw last Tuesday, and you shouldn't report it as if it were.
API answers vs chatgpt.com
Tools collect ChatGPT answers in two ways, and vendors don't always say which they use. Calling OpenAI's developer API is fast and cheap, but the API and the consumer app are separate products with different defaults and different search behaviour. Asking on chatgpt.com itself returns what a visitor gets.
Country creates a similar split. ChatGPT search localizes by where the connection comes from, so a request from a US server gets US sources even when the prompt mentions Spain. If you sell in several markets, confirm that your tool sets the country at connection level. Peec AI documents collection through the interface; many vendors publish nothing, and that alone is a reason to ask.
Reading ChatGPT visibility results
A report is only useful if it changes what you do next. These are the patterns we'd act on:
| What the report shows | What it usually means | Next step |
|---|---|---|
| Low visibility, review sites and listicles cited | Your category's sources don't mention you | Get onto those pages |
| Good visibility, weak tone | ChatGPT repeats a complaint from somewhere | Find the cited source and answer it |
| Your site cited, competitors recommended | Your content is informing answers without selling you | State what you do and who it's for on the page |
| Strong in ChatGPT, weak in Google | The two systems read different sources | Track AI Overviews separately |
| Visibility changes by under 5 points | Probably run-to-run noise | Wait for more runs |
The first row is where most of the work is. We've written up what to do about it step by step in how to rank in ChatGPT.
ChatGPT tracking tools
ChatGPT is the one assistant nearly every AI visibility tool covers, so coverage won't separate them. Price per prompt, collection method and which other assistants come included will. Our best AI visibility tools shortlist sorts them by team and budget, and the Peec AI alternatives comparison goes through plans in detail.
Wherever possible, AI Monitor asks on chatgpt.com in a fresh logged-out session, through a residential connection in your report's country, and flags any answer that came from elsewhere. You get the full text, every source and labels you can correct. It's priced per check, so a weekly set of 20 prompts in ChatGPT alone costs a few dollars a run. Start with free credit and see which of your prompts name you.
FAQ
Can you track your ranking in ChatGPT?
Not as a fixed position, because the list of brands changes between runs. You can track how often you appear across many prompts, which holds up far better.
How do I see if ChatGPT mentions my brand?
Open a temporary chat and ask the questions a buyer would ask, without your brand name in them. Repeat on other days, or use a tracker to do it on a schedule.
Does ChatGPT give everyone the same answer?
No. Answers differ between runs even for the same prompt, and memory, chat history and location add more variation for signed-in users.
How many prompts do I need to track ChatGPT visibility?
Twenty or more per topic is a sensible floor. With five prompts, one answer changes your visibility by 20 points, which makes any trend impossible to read.