How to track brand mentions in ChatGPT, Gemini and Perplexity

Short answer: To track brand mentions in ChatGPT, Gemini and Perplexity, build a list of the questions your buyers ask, run each one in every engine on a regular schedule, and record whether your brand is named, its position, how it is described, which competitors appear and which sources are cited. You can do this by hand in a spreadsheet to start, but because AI answers vary between runs, an AI visibility platform that asks the same questions consistently and keeps the history gives you a far more reliable picture.

There is no Search Console for ChatGPT. No dashboard tells you when an AI engine recommended you, or when it recommended your competitor instead. Yet those answers increasingly decide who makes a buyer’s shortlist. This guide shows you exactly how to track your brand’s mentions across the major AI engines, what to record, and how to tell real movement from noise.

Why tracking AI brand mentions is different

Tracking brand mentions in AI engines is different from social listening or SEO tracking for three reasons: the answers are private, they are personalised, and they change from one run to the next. Each of these shapes how you should track.

• The answers are private. A conversation between a buyer and ChatGPT is not published anywhere, so you cannot monitor it the way you monitor social media or news. You have to ask the questions yourself.
• They are personalised. Answers can differ by account, location, conversation history and settings, so what you see may not be exactly what a buyer sees.
• They are probabilistic. Ask the same question twice and you can get a different answer. A single check is a sample, not a verdict.

The result is that good AI mention tracking looks less like monitoring and more like a controlled survey: the same questions, asked the same way, on a schedule, with the results compared over time.

What to track

Track six things for every buyer question: whether you are mentioned, your position, how you are described, which competitors appear, which sources are cited, and whether the facts about you are accurate.

1. Mention: is your brand named in the answer at all?
2. Position: are you named first, in a short list, or buried below competitors?
3. Description and sentiment: is the engine positive, neutral or negative about you, and what does it say you are best for?
4. Competitors: which other brands are named, and how often do they appear when you do not?
5. Citations and sources: which websites does the engine cite or link to, and is your site among them?
6. Accuracy: are your prices, features, locations and other facts correct? Wrong facts are a common and fixable problem.

How to build your prompt list

Your tracking is only as good as the questions you track. Build a list of 20 to 50 prompts that mirror how real buyers ask, grouped by intent, and keep it stable so your results are comparable over time.

Cover these question types:

• Category questions: “What is the best CRM for a small team?”
• Use case questions: “Which accounting software is best for freelancers in the UK?”
• Comparison questions: “Is Brand A or Brand B better for ecommerce?”
• Alternative questions: “What are the best alternatives to Brand A?”
• Problem questions: “How do I stop customers abandoning their carts?”
• Local questions, if relevant: “Who is the best emergency plumber near Manchester?”

Use your buyers’ language, not your marketing language. Your sales calls, support tickets and site search logs are the best source of real questions. Avoid prompts that include your own brand name, except for a few that check how engines describe you.

Method 1: track mentions manually

You can track AI brand mentions by hand with a spreadsheet. It is slow, but it is free and it teaches you a lot about how engines see your category.

1. Set up a sheet with columns for date, engine, prompt, mentioned (yes or no), position, sentiment, competitors named, sources cited and notes.
2. Reduce personalisation. Use a private or signed out browser session where the engine allows it, and avoid long conversation history that could bias answers.
3. Run every prompt in every engine: ChatGPT, Gemini, Perplexity, Google AI Mode and Google AI Overviews.
4. Run important prompts more than once. Because answers vary, run your key prompts two or three times and record how often you appear rather than a single yes or no.
5. Calculate your mention rate for each engine: the share of prompts where you were named. Then compare it with your top competitors.
6. Repeat weekly on the same day, with the same prompts.

The manual method works for a first look. Its weaknesses are time, consistency and scale: 30 prompts across five engines, run twice, is 300 answers a week to read and log.

Method 2: use an AI visibility platform

An AI visibility platform automates the whole process: it asks your buyer questions across every engine on a schedule, reads each answer, records mentions, position, sentiment and competitors, and turns the results into a score and a trend. The better ones also tell you what to fix.

What automation adds is consistency. The same prompts are asked the same way every time, so a change in your results reflects a real change, not a different phrasing or a different day. ONEGO, for example, tracks ChatGPT, Gemini, Perplexity, Google AI Mode and Google AI Overviews, scores your visibility with the ONE Score out of 100, shows a leaderboard of the competitors named instead of you, keeps the full answer text, and refreshes weekly. You can run a free AI visibility check to see your current mentions before deciding anything. If you are comparing tools, read our guide to the best AI visibility platforms.

Track AI referral traffic in Google Analytics too

Mentions tell you whether AI recommends you. Referral traffic tells you how many people clicked through. Track both, because many AI mentions never produce a click, and AI referrals are easy to miss in your reports.

In Google Analytics 4, create a segment or exploration that filters sessions by source, using a regular expression that matches the main AI assistants:

chatgpt\.com|chat\.openai\.com|perplexity\.ai|gemini\.google\.com|copilot\.microsoft\.com|claude\.ai

This catches visits from people who clicked a link in an AI answer. Some AI traffic still arrives without a referrer and shows up as direct, so treat this number as a floor, not the full picture. It confirms clicks, but only mention tracking tells you about the buyers who saw your name and never clicked.

How often to track, and how to read changes

Track weekly. AI engines update often and competitors move, so weekly checks show the trend without drowning you in noise. Judge your progress on the trend across several weeks, not on any single result.

When you read your results:

• Look for patterns, not points. A drop in one engine for one week may be noise. A steady decline across several weeks is a signal.
• Compare against competitors. If everyone dropped, the engine probably changed. If only you dropped, something about your signals changed.
• Check by engine. Strong in Perplexity but missing from ChatGPT is a specific problem with a specific fix.
• Tie changes to actions. Note when you publish content or earn a mention, so you can see what moved the needle.

What to do when AI gets your brand wrong

If an engine describes your brand inaccurately, fix the sources it reads rather than arguing with the engine. Correct your own website first, then your Google Business Profile, directory listings and other profiles, so the consistent, correct version is what engines retrieve. Add clear Organization schema to your site, and expect the change to take time to show up as engines refresh their sources.

Tracking is the first step, not the last. Once you know where you are named and where you are missing, use our guide to getting your business found in AI search to close the gaps.