How to Track AI Brand Mentions (Manual Method and Tools)

This is the hands-on companion to what AI visibility is.

Why tracking AI brand mentions needs repeated runs

Tracking AI brand mentions needs repeated runs because the same prompt produces different answers. In a SparkToro and Gumshoe study, 600 volunteers ran 12 prompts through ChatGPT, Claude and Google's AI nearly 3,000 times. The chance of two responses listing the same brands was below 1 in 100.

So "ChatGPT mentioned us" is not a measurement. "ChatGPT mentioned us in 11 of 15 runs across five prompts" is. Everything below is built around that rule.

Step 1: Build your prompt set

Your prompt set is the list of questions you'll test every month, and it should mirror what buyers actually ask an AI engine.

1. Collect real questions. Pull them from sales call notes, support tickets, demo requests, Reddit threads and People Also Ask boxes. The Question Finder helps fill gaps.
2. Cover four types. Category ("best CRM for real estate agents"), problem ("how do I stop leads going cold"), comparison ("HubSpot vs Pipedrive"), and brand ("is [your brand] good for small teams").
3. Write them like people talk to chatbots. Full sentences, with context such as team size or budget.
4. Freeze the list. Changing prompts every month breaks comparability. Add new ones in a separate section.

Start with 20 to 30. Fewer is too noisy; more gets hard to run by hand.

Step 2: Choose engines and set up clean sessions

Choose the engines your buyers use and test them in conditions that approximate a new user.

• Engines: ChatGPT, Google (AI Overviews and AI Mode) and Perplexity cover most audiences as of October 2026. Add Gemini, Microsoft Copilot or Claude if your buyers use them.
• Clean sessions: use a logged-out browser or a temporary chat, with memory off, so past conversations don't shape answers.
• Location: keep the same country for every run. Results differ by region.
• Date: run everything within a day or two, so a model update doesn't split your data.

Step 3: Run each prompt and log the results (free template)

Run each prompt three times per engine and log one row per run. Copy these columns into a spreadsheet; this is the whole template.

Twenty-five prompts, three engines and three runs is 225 rows. Budget an afternoon. The "Sources cited" column often turns out to be the most useful: it tells you which third-party pages decide your category's answers.

Step 4: Turn the log into metrics

Turn the log into a few rates you can compare month to month. Calculate each one per engine first, then overall.

1. Mention rate = runs where you were mentioned ÷ total runs.
2. Citation rate = runs where your domain was cited ÷ total runs.
3. Share of voice = your mentions ÷ all mentions of you and your tracked competitors.
4. Top-three rate = runs where you were among the first three brands named ÷ total runs.
5. Sentiment mix = share of mentions that were positive, neutral or negative.

Ignore position averages from small samples, and don't react to a few points of movement between months. The reasoning behind each metric is in AI visibility metrics, and the share calculation is covered in detail in AI search share of voice.

Add first-party data

First-party data confirms whether AI mentions turn into visits. Three sources are worth adding:

• ChatGPT referrals. ChatGPT adds utm_source=chatgpt.com to many outbound links, so visits show up in analytics as referral traffic.
• Bing Webmaster Tools. Microsoft's AI Performance report, in public preview since February 2026, shows how often your pages are cited in Microsoft Copilot and Bing AI summaries.
• Google Search Console. Google's AI features documentation says AI Overviews and AI Mode traffic is included in the Web search type of the Performance report. As of October 2026 it isn't broken out separately.

When to switch to an AI visibility tracking tool

Switch to an AI visibility tracking tool when the manual method stops being repeatable. Clear signs:

• You need weekly or daily data, not monthly.
• You track more than about 50 prompts or more than three engines.
• Several people need dashboards or exports.
• You want prompt ideas from a database rather than your own list.

Tools run your prompts on a schedule, often several times per check, and calculate the rates for you. We compare them by engines covered and metrics in the best AI visibility tools. Keep your manual prompt list: it's the best input to any tool. Once tracking is running, AI brand visibility covers how to move the numbers.

Frequently asked questions

What is an AI visibility tracking tool?

An AI visibility tracking tool runs a set of prompts through AI engines like ChatGPT, Perplexity and Google AI Overviews on a schedule, then reports how often your brand is mentioned, which sources are cited, how you're described, and how you compare with competitors. Most use prompts you supply. Some add prebuilt prompt databases with estimated volumes.

Can I track AI brand mentions for free?

Yes. Use the manual method: 20 to 30 buyer prompts, three engines, three runs each, logged in a spreadsheet with columns for mention, position, citation, sentiment, competitors and sources. It takes an afternoon a month. Add free first-party data from ChatGPT referral traffic in your analytics and the AI Performance report in Bing Webmaster Tools.

How many times should I run each prompt?

Run each prompt at least three times per engine for a manual check, and more if a tool does it for you. AI engines rarely return the same brand list twice, so a single run can mislead badly. Three runs across 20 to 30 prompts gives a usable baseline; trends across several months matter more than any single snapshot.

How often should I track AI brand mentions?

Monthly is enough for a manual program, using the same prompts each time so results stay comparable. Run an extra check after a major site change, a product launch, a press push or a big model release. If you need to see the effect of specific campaigns quickly, move to a tool that runs prompts weekly or daily.

If you want this tracking set up and acted on for you, AEO GEO Labs runs GEO, AEO and SEO programs for B2B and SaaS teams. See our services.