How to Track Brand Mentions in ChatGPT and AI Overviews

Quick answer: Start with the two free sources you already have. Search Console's Generative AI report shows how often your pages appear in Google AI Overviews and AI Mode, and GA4's AI Assistant channel shows visits from ChatGPT, Gemini, Copilot and other assistants. Neither shows the conversations where you were recommended and nobody clicked, or where a competitor was recommended instead. To see those, run a fixed set of buyer prompts against each model on a schedule and record how often you appear, where, and which sources the model cited. You can do that by hand in an afternoon, or pay $29 to $250 a month for a tool that runs it daily.

Below is the method we run on ourselves, including the finding that changed what we worked on.

Why this became worth measuring

A meaningful share of product research now happens inside a chat window. Someone evaluating tools asks ChatGPT for the best options, gets five names, and looks into two of them. If you are not in the five, you were never in the consideration set, and nothing in your analytics records that it happened.

When you lose a Google ranking you can see it. When you lose an AI recommendation you see nothing: no impression, no position, no referrer. The traffic simply never arrives.

That is the reason to measure it deliberately.

What you can see for free

Before building anything, check what Google already reports. It covers more than it did a year ago.

Google AI Overviews and AI Mode

Search Console added a Generative AI performance report in June 2026 and rolled it out to every site on August 31, 2026. It counts how often a link to your site appeared inside AI Overviews or AI Mode, and lets you break that down by page, country and device.

It does not include clicks or queries. You can see that a page was shown in AI answers 4,000 times last month, but not what people asked. If the report is missing from your property, Google says the usual reason is that the site has not had enough AI-feature impressions yet.

Clicks from AI Overviews and AI Mode land in GA4 as Organic Search. Google's channel definitions say so explicitly, so you cannot separate them from regular organic clicks in GA4.

ChatGPT, Perplexity, Gemini, Claude and Copilot

GA4's default channel group now includes an AI Assistant channel for visitors arriving from recognized assistants. Google names ChatGPT, Gemini, DeepSeek, Copilot and Grok as examples. Check it under Reports, then Acquisition, then Traffic acquisition. It was added in May 2026 and does not reclassify older data, so earlier AI visits still sit in Referral or Direct.

ChatGPT adds utm_source=chatgpt.com to links it cites, which GA4 uses as the session source. Because there is no matching utm_medium, some of those sessions can end up as Unassigned. If you want full control over which assistants count, create a custom channel group with a channel called "AI assistants", set Source to match this regex, and move it above Referral:

^(.*\.)?(chatgpt\.com|chat\.openai\.com|perplexity\.ai|gemini\.google\.com|claude\.ai|copilot\.microsoft\.com|grok\.com|chat\.deepseek\.com)$

Treat the result as a floor. Clicks from the ChatGPT apps and some browsers arrive with no referrer and get counted as Direct.

What neither tool can show you

Both sources only see the conversations that ended in a click on your link. They cannot tell you how often you were recommended without a click, how often a competitor was recommended instead, or what the model says your product does. For that you have to ask the models yourself.

Why prompt tracking is harder than SEO tracking

Four properties make AI mentions difficult to measure. Any tool or method claiming precision here is overselling.

Answers are non-deterministic. In a study by SparkToro and Gumshoe, 600 volunteers ran 12 recommendation prompts through ChatGPT, Claude and Google's AI 2,961 times. There was less than a 1 in 100 chance of getting the same list of brands twice, and closer to 1 in 1,000 of getting it in the same order. Even the length of the list changed from run to run. A single check tells you almost nothing.

There is no official surface for chat assistants. Google now reports AI Overview impressions. OpenAI, Anthropic and Perplexity give publishers nothing comparable, so everything about them is inferred from the outside.

Personalization varies. Memory, custom instructions and account history all shape answers. Your own account is the worst place to test, because it has been reading your site all year.

Models update without notice. A result from March may not reproduce in July, and nothing announces the change. Trends matter more than any single measurement.

The method

Step 1: Build a prompt set that reflects real buying behavior

This step determines everything downstream, and it is the one people rush.

Write 20 to 40 prompts in four groups.

Category prompts. What someone asks before they know any brands.

• "best AI marketing tools for a solo founder"
• "how do I do marketing without hiring anyone"

Comparison prompts. Mid-evaluation.

• "Okara vs NoimosAI"
• "alternatives to hiring a marketing agency"

Problem prompts. The buyer describes a situation instead of a category.

• "I have a SaaS product and no time for marketing, what should I use"
• "how do I get my startup mentioned on Reddit without getting banned"

Brand prompts. What the models say when asked about you directly.

• "what is Okara"
• "is Okara any good"

Problem prompts are the ones most people skip, and they matter most. Buyers describe symptoms far more often than they name categories, and the brand recommended in response to a symptom is the one that enters the consideration set.

Brand prompts serve a different purpose. They tell you whether the models understand what you sell at all. For us, that turned out to be the more urgent finding.

Step 2: Query the models cleanly

Run every prompt against each model you care about. For most B2B products that means ChatGPT, Perplexity, Google AI Mode, Gemini and Claude, though the list keeps changing.

Three rules matter more than they sound:

Use the API or a clean session, not your own account. App sessions carry memory and personalization. API calls with no system prompt are the closest you get to a neutral observer.

Run each prompt many times. Five runs per prompt is a workable floor for a first look. SparkToro's researchers suggest 60 to 100 runs before you treat a list as a stable picture. You are measuring how often you appear, not whether you appeared. "We appear in 3 of 5 runs" is real information. "We appeared" is not.

Log the full response, not a yes or no. The competitors named alongside you and the sources cited are more actionable than your own mention.

Step 3: Record four things per run

A spreadsheet with one row per run and columns for date, model, prompt, mentioned, position, framing, competitors named and cited URLs is enough to start.

The citation column is where the leverage is. If a model recommends three competitors and cites the same two listicles each time, you know exactly which pages to get into. That is a concrete task, unlike "improve our GEO."

Step 4: Automate it

Manual checks work for a first pass and then stop scaling. We built ours on the DataForSEO LLM mentions API running on a schedule, writing to a database, with a dashboard showing mention rate over time per prompt per model.

If you would rather not build, these tools run the same loop for you. List prices were checked in October 2026, billed monthly.

Every vendor counts something different (prompts, checks, credits or answers), so the cheapest sticker price is not always the cheapest per prompt tracked. For a longer comparison, see our guide to the best GEO tools.

What we found, including the part we did not expect

Three things came out of running this on ourselves.

Brand prompts exposed a problem the category prompts hid. We were tracking whether we appeared for "best AI marketing tools." We were not tracking what models said when asked "what is Okara." Several described us as a private AI chat product with encrypted conversations and open-source model access.

That is not what we sell. It was what we used to sell, and the models had learned it from third-party directory listings that were never updated, plus legacy pages on our own site.

No amount of work on category prompts would have fixed that, because the problem was entity definition rather than visibility. If the models do not know what you are, ranking for your category is not the constraint.

Check your brand prompts before anything else. If the answer to "what is [your product]" is wrong, fix that first. Directory listings, your own legacy content and schema markup are the levers, and none of them are content marketing.

Cited sources clustered far more than expected. Across dozens of runs, the same handful of roundup articles kept appearing as citations. Vendor pages, including ours, appeared far less often than third-party comparisons. Getting included in other people's listicles moved the needle more than publishing our own.

Consistency across models is low. A brand that shows up often in Perplexity can be invisible in ChatGPT. Perplexity leans on live retrieval, so recent content and clean citation structure help. ChatGPT leans more on what it learned in training, so long-standing presence and third-party mentions matter more. These need different work.

What to do with the data

If your mention rate is near zero on category prompts: the constraint is third-party presence, not your own content. Get into the roundups your competitors are cited in. Your citation log tells you which ones.

If you are mentioned but framed poorly: the models have a view of you that is unflattering or too narrow. This usually traces to how comparison sites and directories describe you, not your own site.

If your brand prompts return the wrong description: stop other GEO work and fix the entity. Update directory listings, add unambiguous schema, publish one clear page that says what you do, and remove or de-index legacy content describing a product you no longer sell. This is the highest-leverage GEO work available, and almost nobody does it because it does not look like marketing.

If you are strong in one model and absent in another: treat them as separate channels. Perplexity rewards fresh, well-cited content. ChatGPT rewards accumulated third-party presence. Our GEO checklist covers the fixes for both.

Frequently asked questions

How often should I check AI brand mentions?

Weekly is enough for most teams. Daily checks mostly produce noise, because answers change from run to run anyway. Monthly is too slow to catch a model update.

Can Google Search Console track AI Overviews?

Partly. The Generative AI performance report shows how often your pages appeared in AI Overviews and AI Mode, broken down by page, country and device. It does not show clicks or the queries that triggered the answer, and it covers nothing outside Google.

Does traffic from ChatGPT show up in Google Analytics?

Yes, when someone clicks a link. GA4 groups recognized assistants into the AI Assistant channel, and ChatGPT tags its links with utm_source=chatgpt.com. Some app clicks arrive without a referrer and show up as Direct, so treat the number as a minimum. What analytics cannot show is the much larger set of conversations where you were recommended and nobody clicked.

Is there a free way to track AI brand mentions?

Yes. Use Search Console's Generative AI report for Google, the GA4 AI Assistant channel for chat assistants, and a manual prompt run for everything else: 20 prompts, run 5 times each across 3 models, logged in a spreadsheet. That is 300 queries and an afternoon, and it will tell you more than most GEO audits. Move to a tool once you want a trend rather than a snapshot.

What is a good AI mention rate?

There is no benchmark, and anyone quoting one is guessing. Track your own trend and your rate relative to named competitors on the same prompts. Relative position is the only meaningful number here.

Which models should I track?

ChatGPT and Google's AI Overviews and AI Mode at minimum, since they reach the largest audiences. Add Perplexity if your buyers are research-heavy, and Claude if you sell to developers or technical buyers.

Does llms.txt affect whether models mention me?

Evidence is thin so far. It is cheap to add and unlikely to hurt, but treat claims about its impact skeptically. Entity clarity and third-party citations demonstrably matter more. Our llms.txt guide covers what it does and does not do.

Sources

• Generative AI performance report, Google Search Console Help
• Default channel group, including AI Assistant and Organic Search definitions, Google Analytics Help
• AIs are highly inconsistent when recommending brands or products, SparkToro and Gumshoe
• How much do AI visibility tracking tools cost in 2026?, Media Morpheus, prices read October 19, 2026
• Vendor pricing pages: Otterly, Peec AI, Semrush, Ahrefs Brand Radar, Scrunch, Profound