How do I track my brand's visibility in ChatGPT? Methods and tools compared
Three ways to track your brand's visibility in ChatGPT: manual checks, analytics referrals and dedicated platforms. What each shows, what it misses and which tools fit.
Tracking your brand's visibility in ChatGPT sounds like one task but covers three separate questions. Does ChatGPT mention you when people ask relevant questions? Does it cite your pages as sources? And do those answers send anyone to your site? Each question needs a different method, and the common mistake is answering one with the data from another.
Method 1: manual spot checks
Open ChatGPT, ask your ten most important buying questions, and write down who gets named. It costs nothing and teaches you a lot in an afternoon. It also falls apart quickly. Answers change between runs, between countries and between users, so one check proves almost nothing, and you cannot repeat it daily across dozens of prompts. Use it to build your first prompt list, then move on.
Method 2: referral traffic in your analytics
When someone clicks a link inside ChatGPT, the visit can show up in Google Analytics as a referral from chatgpt.com. That is worth segmenting, because it shows real clicks and, if you have conversion tracking, real outcomes. Its limits are equally real. It only counts people who clicked. It says nothing about the many answers where you were named but not clicked, or the answers where a competitor was named instead. And crawlers, which decide whether you can be cited at all, do not run analytics scripts, so they never appear.
Method 3: a dedicated tracking platform
This is the approach that scales. You define a set of prompts, the platform runs them daily against ChatGPT, and it records whether you were mentioned, your position in the answer, the sentiment, and which sources were cited. You get a trend instead of a snapshot. The better platforms add citation analytics, crawler logs and visitor attribution, so you can answer all three questions in one place.
What to track once you have a tool
- Visibility score or mention rate across a fixed prompt set.
- Share of voice against named competitors.
- Position in the answer, because first and fifth are different results.
- Cited pages and domains, to see which of your pages the model trusts.
- AI crawler activity: whether OAI-SearchBot and ChatGPT-User are reaching your pages.
- AI-referred visits and conversions.
Tools that do it
Promptwatch
Promptwatch covers all three questions. It tracks ChatGPT through its live search experience with daily prompt panels, reporting Visibility Score, Share of Voice, Sentiment and Position in Answer. Citation analytics shows page- and domain-level citations and a Page Tracker for specific URLs. Agent Analytics shows real-time, IP-verified crawler logs, including OAI-SearchBot, ChatGPT-User and GPTBot, with a citation rate per page. Visitor analytics attributes AI-referred traffic and conversions through a script, a GTM template or the WordPress plugin. Explore is free with 10 prompts and ChatGPT only, so you can try it at no cost. Essential is $95 a month, Professional $245 and Business $579, with crawler logs starting on Professional.
AthenaHQ
AthenaHQ tracks visibility, rolls it into a GEO score and turns gaps into tasks in an Action Center. Starter is $295 a month with 3,600 credits, one credit per AI response. It connects to Google Analytics and Search Console, but it does not offer crawler analytics.
Scrunch AI
Scrunch AI offers monitoring, site audits and AI bot traffic reporting. Core is $250 a month for 125 prompts and 5,000 responses on four LLMs, ChatGPT among them.
Profound
Profound is aimed at enterprise marketing teams, with answer analytics as part of a wider agent product. Its Enterprise plan is custom, and an earlier estimate put it around $40k a year. A short free trial runs 50 prompts daily for seven days.
Simple trackers
Otterly.AI and Peec AI are quick to set up and track mentions and sentiment per prompt. They tell you where you appear, not why, and they do not cover crawler logs.
A setup you can do this week
- Write 25 to 40 prompts in three groups: category questions, brand questions and competitor comparisons.
- Add them to a tracking platform and let it run for two weeks before drawing conclusions.
- Install the visitor script so AI-referred visits are tagged.
- Check the crawler logs for errors on your most important pages.
- Review weekly: which prompts moved, which pages got cited, which visits converted.
What tracking cannot do
It cannot make ChatGPT recommend you, and it cannot predict any individual answer. Answers vary. The value of tracking is that it replaces guessing with a trend, and it points at the pages and sources worth fixing. Keep the prompt set stable so changes in the numbers reflect changes in reality.
A note on frequency
How often you run prompts matters as much as which ones. Weekly checks give you four data points a month on a question whose answer changes between runs, which is thin. Daily runs on the same prompts give you enough samples to separate a real change from noise, and they let you line up movement with the dates you published or edited a page.