Tracking Your Brand on ChatGPT Claude Gemini and Perplexity

TL;DR: how do you track your brand in ChatGPT, Claude, Gemini and Perplexity?

To track your brand in ChatGPT, Claude, Gemini and Perplexity, build a fixed list of buyer prompts, run them on every engine on a schedule, and record whether you are mentioned, cited, ranked and described accurately. Do it by hand for a first baseline, then automate. Listable Labs automates ChatGPT, Gemini and AI Mode from $89/month, but not Claude.

The steps below work with a spreadsheet and an hour a week. The tool section shows where automation pays off and which tools cover Claude.

Who this guide is for and how it was put together

This guide is for marketers, founders and agencies who want to know whether AI engines mention their brand, and what to do if they do not. It is organised as the steps I would give a new client, judged on four questions:

• Can a beginner do it without paid software?
• Does it produce numbers you can compare week to week?
• Does it show which sources drive the answer?
• Where does it stop scaling and need a tool?

Tool details and prices were checked on vendor pages in October 2026.

What does each engine need from a manual check?

Each engine behaves differently, so check them separately and keep the notes side by side.

Step 1: How do you build the prompt list?

Write 20 to 50 prompts that a real buyer would type, because tracking a brand name alone only tells you what AI says when someone already knows you. Cover five types:

• Category: "best project management software for agencies"
• Comparison: "Brand A vs Brand B"
• Alternatives: "alternatives to [competitor]"
• Use case or persona: "CRM for a five-person consultancy"
• Problem: "how do I track brand citations in ChatGPT?"

Add your market and city where it matters. Keep the list fixed once you start, otherwise week-to-week changes will reflect your edits, not the engines.

Step 2: What should you record for each answer?

Make a sheet with one row per prompt, engine and date. Record the seven metrics below, which are the same ones paid tools report:

1. Mention: is your brand named at all?
2. Citation: does the answer link to or cite one of your pages?
3. Position: where do you appear in a list of recommendations?
4. Competitors: which rivals are named, and in what order?
5. Sources: which URLs and domains does the answer cite?
6. Sentiment: is the description positive, neutral, comparative or negative?
7. Accuracy: are the price, features and facts correct?

From these you can calculate mention rate (answers that name you divided by answers run), citation rate and share of voice (your mentions divided by all brand mentions).

Step 3: How do you run a clean manual check?

Answers vary with chat history, memory, location and account settings, so remove as many variables as you can.

1. Open a new chat or a private window for every prompt.
2. Turn off memory and custom instructions where the engine allows it.
3. Use the same engine mode each time, for example with web search on.
4. Run each prompt three times and record all three, because single runs mislead.
5. Save the full answer and a screenshot with the date.

Expect noise. A brand named in two of three runs is meaningfully different from one named in none, and that rate is more useful than any single result.

Step 4: How do you read the sources behind an answer?

The cited sources explain why a brand was chosen. For every prompt where a competitor is named and you are not, open the cited URLs and note the type: comparison article, review site, forum thread, publisher guide or the competitor's own page.

That list tells you where to act. If most citations are list articles you are missing from, you need outreach. If they are forum threads, you need community presence. If they are your competitors' own pages, you need better pages. Perplexity and AI Mode show citations clearly; ChatGPT and Claude show them when web search is used.

Step 5: How do you score sentiment and accuracy?

Being mentioned is not enough if the answer is wrong or lukewarm. Classify each mention as positive (recommended), neutral (listed without comment), comparative (positioned against others), negative (warned against) or absent.

Then check the facts. Wrong pricing, discontinued products and mistaken locations are common, and they are fixable: update the pages and profiles the engine cites, and make the correct fact easy to find. Keep a log of errors so you can confirm they disappear.

Step 6: When should you move from a spreadsheet to a tool?

Manual tracking breaks at about 30 prompts across four engines, run three times a week: that is hundreds of answers to read and log. Move to a tool when you need history, daily runs, competitor benchmarks or reports for a client.

Look for these features:

• Engines that match where your buyers search, including Claude if relevant.
• A prompt limit and answer volume that fit your list.
• Cited-source data, not only a mention count.
• Competitor benchmarking and share of voice.
• Exports or white-label reports if you serve clients.

Which tools can track your brand across these engines?

Prices below were checked on vendor pages in October 2026.

Listable Labs tracks ChatGPT, Google Gemini and Google AI Mode on every plan, with Perplexity, Google AI Overviews and Microsoft Copilot as add-ons. It shows mentions, citations, share of voice and competitor benchmarks, and connects them to GA4, Search Console, Bing Webmaster Tools and Cloudflare data. Growth is $89/month for 50 prompts and 4,500 analysed answers, Scale is $249/month for agencies, and Max is $400/month. It does not track Claude, so pair it with manual Claude checks. See Listable Labs plans.

For Claude coverage, these tools list it among their engines:

• **Peec AI:** from $95/month for 50 prompts and a choice of three models; more models on Enterprise.
• **Promptwatch:** from $95/month for 50 prompts across four models.
• **LLMrefs:** $79/month for 500 prompts across all engines.
• **Trakkr:** from $100/month for 50 prompts across eight models.

Claude is an add-on at some vendors, so confirm it is in your plan. For full tool comparisons, read the best brand AI visibility tracking tools guide, the ChatGPT tracking tools roundup and the pricing and coverage comparison.

How do you turn tracking into action?

Tracking only matters if it changes what you publish. Work through the gaps in this order:

1. Fix inaccurate facts on your own pages and profiles.
2. Publish answer-ready pages for prompts where rivals appear and you do not.
3. Pursue mentions on the sources that engines already cite.
4. Join community discussions that engines quote.
5. Re-run the same prompts weekly and compare against your baseline.

If you use an MCP-compatible assistant, the Listable MCP server lets you query visibility, mentions, citations and competitors from Claude, ChatGPT, Cursor and other clients, which is a handy way to check results inside your daily tools.

Tracking checklist

• A fixed list of 20 to 50 buyer prompts across five types.
• A sheet or tool recording the seven metrics per prompt, engine and date.
• Three runs per prompt for manual checks, in clean sessions.
• A source log for every prompt where you are missing.
• A sentiment and accuracy score for every mention.
• A plan for Claude coverage, manual or via a tool that supports it.
• A weekly review and a monthly summary of mention rate and share of voice.

Verdict

Start by hand so you understand what the engines say about you, then automate once the list is stable. Use a tool that covers the engines your buyers use, and fill any gap, such as Claude with Listable Labs, with manual checks or a second tool.

To automate the baseline, start free on Listable Labs with no credit card, or ask for a free discovery audit if you want the team to run it for you.