How to Track Brand Mentions in ChatGPT (and Other AI Assistants)

When a buyer asks ChatGPT "what's the best tool for X?", your brand either shows up in the answer or it doesn't. Unlike Google, there is no rank tracker built in, so most teams have no idea which side they're on.

This guide covers how to measure it properly.

Why manual checks mislead you

The obvious approach is to open ChatGPT and type your brand name. It fails for three reasons:

• Asking about your brand by name is a rigged test. The model will talk about you because you put your name in the prompt. What matters is whether it recommends you when the buyer doesn't name you.
• Answers vary between runs. The same prompt can produce different brands on different days, so a single screenshot proves nothing.
• Each model behaves differently. A brand that shows up in Perplexity (which cites live web sources) can be invisible in a model answering from training data alone.

What to measure

Three metrics cover most of what you need:

Metric	Question it answers
Mention rate	In what share of relevant questions is the brand mentioned at all?
Ranking position	When mentioned, is it first, third or tenth in the list?
Sentiment	Is the mention positive, neutral or negative?

Track them per platform, not just as one blended number.

Build a question set that reflects real buyers

Write questions the way your customers ask them, in three groups:

1. Category questions: "What are the best project management tools for agencies?"
2. Comparison questions: "Compare Brand A and Brand B for small teams."
3. Problem questions: "How do I keep client work organized across five projects?"

Only the first and third groups tell you whether the model brings up your brand unprompted. Keep direct brand questions in a separate bucket so they don't inflate your numbers.

Repeat it on a schedule

Because answers vary, run the same set weekly and look at the trend, not a single run. A brand whose mention rate moves from 20% to 35% over two months is improving; one that bounces between 25% and 30% is stable.

Tip: PeiGEO automates this loop: it generates a tracking question set from your brand and competitors, runs it across DeepSeek, Qwen, ChatGPT, Gemini and Perplexity, and reports mention rate, ranking and sentiment over time.

What to do with the results

If the brand is missing, look at what the models do cite. Perplexity and similar tools show their sources, and those pages (review sites, comparison articles, forums) are where your effort should go. See our user guide for how the metrics are calculated.