SiteSignal
To track brand mentions in ChatGPT, write 10 to 20 questions buyers ask and run each in a new chat every week. For every answer, log whether your brand is named, its position, which sites are cited, which competitors appear and whether the facts are right. Repeat them in Gemini and Perplexity, and compare weeks, not single answers.
How do I check if ChatGPT mentions my brand?
Open a new chat in ChatGPT with memory off, or use a Temporary Chat, so your past conversations don't steer the answer. Ask a question a buyer would ask, such as "What are the best [your category] providers for [your audience]?". Read the whole answer, not just the first line. Note whether your brand appears, where it sits in the list, how it is described and which competitors are named instead. Then ask a direct question about your own facts, such as "What does [your brand] charge?" or "Where is [your brand] based?". That second question tells you whether ChatGPT knows you at all, and whether it knows you correctly.
Which questions should I track?
The questions you choose decide what you see. In SiteSignal's UK workwear research, 99 to 100% of answers to "which supplier" questions named at least one supplier. Answers to "how do I choose" questions named one in only 0 to 6% of cases (22 August to 21 October 2026, State of AI Visibility in UK Uniform & Workwear). So if you only track advice questions, you will see almost no brands at all. A balanced set has four kinds of question:
- Category lists: "Best [category] for [audience]", "Top [category] in [location]".
- Comparisons: "[Your brand] vs [competitor]", "Alternatives to [competitor]".
- Problems: "How do I [solve the problem you solve]?"
- Your facts: prices, locations, contact details, opening hours, policies.
Keep the wording identical every time you run them. A small change in wording can change the answer, and then you can't compare one week with the next.
What should I record for each answer?
Log one row per question, per engine, per date. These columns cover what most teams need:
| Column | What to write | Why it matters |
|---|---|---|
| Mentioned | Yes or no | The base of your inclusion rate |
| Recommended | Yes if the answer suggests you, not only lists you | A mention and a recommendation are not the same win |
| Position | 1st, 2nd, 3rd and so on | Buyers read the top of the list first |
| Cited URL | Any link or source the answer gives | Shows which pages AI trusts in your category |
| Facts right | Right, partly right, no answer or wrong | Catches hallucinations about your business |
| Competitors named | Every other brand in the answer | The base of your share of mentions |
Save the full answer text as well. When a number moves, the text tells you why.
How do I track brand citations in AI answers?
A citation is a source an AI answer links to or names as evidence. A mention is your brand name in the answer text. They measure different things. An answer can cite your guide while recommending a competitor. We learned this in our own research. Until 21 August 2026 our system recorded the websites an answer linked to or named. From 22 August it recorded the brand names written in the answer itself. The two measure different things, so we don't compare numbers across that date. So log citations in their own column. When ChatGPT searches the web, it shows its sources under the answer. Note the domain and the exact page each time. Over a few weeks you will see which sites AI relies on in your category. Those are the pages worth getting listed on, reviewed on or quoted by.
How do I track brand mentions in Gemini and other AI models too?
Use the same questions, the same wording and the same columns in each engine. Then compare them side by side, because engines behave very differently. In 6,199 fact-checks on UK workwear brands from May to October 2026, Gemini gave no specific answer 65.3% of the time. ChatGPT gave no specific answer 21.9% of the time, and Perplexity 20.6%. Perplexity was wrong most often, at 8.5% of checks. A brand that looks fine in ChatGPT can be invisible in Gemini. Google AI Overviews is worth adding too, because it appears on top of normal Google results. Our guide to tracking brand mentions in Gemini covers Gemini's quirks.
Which metrics show progress over time?
Four numbers are enough to start. Inclusion rate is the share of answers that name you. Share of mentions (sometimes called share of model) is your mentions divided by all brand mentions in those answers. Average position is where you sit when you are named. Accuracy is the share of fact questions answered correctly.
Here is a worked example of share of mentions. Across 100 answers, Competitor A is named 78 times, you 45 times and Competitor B 34 times. That is 157 brand mentions in total. Your share is 45 divided by 157, or 28.7%. Track the trend over weeks rather than chasing a single figure, because individual answers vary from one run to the next.
How often should I check ChatGPT for brand mentions?
Weekly is the minimum for a manual routine, and it gives you comparable data. Daily is better if you can automate it, because short drops are easy to miss. The Uniform Consultants shows why. During a website rebuild in June and July 2026, Gemini stopped naming the company completely. A monthly check could have missed the whole episode. SiteSignal runs every tracked question once a day for this reason, and it pauses a question once its answer has stayed the same on every engine for 14 days.
When does manual tracking stop working?
It stops when the routine slips. 20 questions in 4 engines is 80 answers to read and log each week, before you look at a single competitor. Most teams manage that for a month or two. After that, gaps appear in the data and the trend line stops meaning anything. That is the point to automate. If you are comparing options, our roundup of the best tools to track brand mentions in ChatGPT lists eight, with prices.
SiteSignal (our product) asks your buyers' questions every day in ChatGPT, Gemini, Perplexity and Google AI Overviews. It scores every answer on an 8-level scale from "Recommended + Cited" to "Not Visible", logs every cited domain and URL, and fact-checks your saved brand facts daily. Each week it turns the gaps into a ranked action plan, split into developer and content tasks.
A real result: The Uniform Consultants
The Uniform Consultants (TUC), an independent UK uniform consultancy, tracked five to six buyer questions every day in ChatGPT, Perplexity, Gemini and Google AI Overviews from 16 May to 23 September 2026.
| May 2026 | September 2026 | |
|---|---|---|
| Share of all brand mentions | 7.3% | 13.6% |
| Average position when named | 2.4 | 1.0 |
| AI engines naming TUC | 2 of 4 | 4 of 4 |
During the June and July rebuild, the average position slipped to 2.9 and Gemini dropped TUC. Once the new site settled in August, every engine named TUC first. In September it led 185 named brands; the next had 3.6% of mentions. For a wider view of one industry, see how AI answers UK workwear buyers.
A 4-week plan to start
- Week 1, baseline: write 10 to 20 questions across the four kinds above. Run them in ChatGPT and one other engine. Work out your inclusion rate and share of mentions.
- Week 2, facts: list your key facts (prices, locations, contact details). Ask each engine about them and mark each answer right, partly right, no answer or wrong.
- Week 3, sources: review the cited URLs. Find the pages AI trusts where you are missing or described wrongly.
- Week 4, act and re-check: fix the wrong facts on your own site first, then work on the sources. Run the same questions again and compare.
Frequently asked questions
How do I check brand mentions in ChatGPT for free?
Open a new chat with memory off, ask the questions your buyers ask, and record whether your brand is named, its position, the sources cited and the competitors named. Repeat the same questions every week in a spreadsheet so you can compare results.
How do I track brand mentions inside AI models like ChatGPT or Gemini?
Run the same fixed questions in each engine on a schedule and log each answer in the same columns. Compare engines side by side, because they behave differently. In our workwear data, Gemini gave no specific answer 65.3% of the time, against 21.9% for ChatGPT.
How do I track brand citations in AI?
Record the sources each answer links to or names, separately from brand mentions. Note the domain and the exact page. After a few weeks you will see which sites AI relies on in your category, and which of them leave you out.
Can you track ChatGPT brand mentions in real time?
No. ChatGPT does not publish the answers it gives other people. You can get close by asking a fixed set of questions every day and watching for sudden changes in mentions, position or accuracy.
What is share of model?
Share of model is your share of all brand mentions across a fixed set of AI answers, like share of voice in advertising. If you are named 45 times out of 157 brand mentions, your share is 28.7%.
Why does ChatGPT give different answers to the same question?
ChatGPT writes each answer fresh, and it may search the web for some questions and not others. That is why one answer proves little. Track the same questions over weeks and look at the trend.
See what AI says about your brand
Get every buyer question checked daily in four AI engines, with fact-checks and a weekly action plan. Start your 7-day free trial. No card needed.