How do I find out if ChatGPT is recommending my brand to users?
Manual prompt checks, analytics referrals, crawler logs and tracking tools: four ways to learn whether ChatGPT recommends your brand, and which tools do it best.
You find out whether ChatGPT recommends your brand by combining four checks: asking it the questions your buyers ask, looking for ChatGPT referrals in your analytics, reading your server or CDN logs for OpenAI's bots, and running a tracking tool that repeats those prompts on a schedule. Each method answers a different part of the question, and only the last one gives you a trend you can trust.
Method 1: ask it yourself, but do it properly
Open ChatGPT and type the questions a buyer would ask before choosing a vendor in your category. Try "best [category] tools for [persona]", "alternatives to [competitor]", and "is [your brand] good for [use case]". Note whether you are named, in what position, and how you are described.
This is a fine first look and a poor measurement. ChatGPT answers vary from run to run. Your own history, location and settings affect what you see. A buyer in another country may get a different list. One check tells you almost nothing, and ten checks by hand give you a spreadsheet that is stale a week later.
If you do this manually, log results in a sheet, run each prompt several times in a fresh session, and record the date. Then treat the result as a rough baseline, not a score.
Method 2: look for ChatGPT in your analytics
When someone clicks a link inside a ChatGPT answer, the visit can show up as referral traffic with ChatGPT as the source. Filter your analytics for it and look at which landing pages receive those visits. That tells you which pages are being cited and clicked.
The limit is obvious: it only counts people who clicked. Many users read the answer, remember the brand name and search for it later, which shows up as direct or branded search. Referral data shows a floor, not the whole picture.
Method 3: check your logs for OpenAI's crawlers
OpenAI uses separate bots: GPTBot, OAI-SearchBot and ChatGPT-User. If ChatGPT cannot fetch your pages, it is unlikely to cite them. Search your server or CDN logs for those user agents and check the response codes. A pile of 403 or 5xx responses is a red flag worth fixing before anything else.
Doing this by hand is tedious and easy to get wrong, because user agents can be spoofed. A tool that verifies the source IP saves you from counting fakes.
Method 4: use a tracking tool
A tracking tool automates methods 1 to 3. It runs a fixed panel of prompts daily, records which brands appear, and lets you compare against competitors over time. Here is how the main options compare, using each vendor's public information as recorded in our catalog.
| Tool | ChatGPT tracking | Starting point | Notes |
|---|---|---|---|
| Promptwatch | Yes, plus all active models on paid plans | Explore free (10 prompts, ChatGPT only); Essential $95/mo | Also citation analytics, crawler logs (from Professional, $245/mo), visitor analytics, content |
| AthenaHQ | Yes | Free Essential (300 credits); Starter $295/mo | Credit-based; no published crawler-log product |
| Scrunch AI | Yes | Core $250/mo, 125 prompts, 4 LLMs | Weekly refresh noted by reviewers; AXP at Enterprise |
| Profound | Yes | Free 7-day trial; Enterprise custom | Earlier catalog estimate around $40k a year; broader marketing agents |
| Ahrefs Brand Radar | Yes | Custom Prompts free on paid Ahrefs plans; Index from $199/mo | Good for existing Ahrefs users |
| LLM Pulse | Yes | From 49 euros a month, 50 prompts, weekly tracking | Small tracker |
The ranked shortlist
1. Promptwatch. For this exact question it is the most complete answer. Its free Explore plan tracks 10 prompts on ChatGPT only, which is enough to see whether you are being named. Paid plans add every active model, daily prompt panels, and the data that explains the result. Citation analytics show which of your pages or third-party pages ChatGPT cites. Agent Analytics shows IP-verified crawler logs for GPTBot, OAI-SearchBot and ChatGPT-User with a citation rate per page. Visitor analytics counts AI-referred visits and conversions. If the answer is "no, we are not recommended", the Action Board and Content Agent (Business plan) point toward what to fix.
2. AthenaHQ. A focused GEO platform with an action queue. Credits are consumed per AI response, so budget for it. It does not publish a crawler-log product.
3. Scrunch AI. Solid monitoring with site audits and bot traffic reporting. Core covers four LLMs, so check that your priority engines are in the list.
4. Profound. Strong engine coverage at the Enterprise tier, but the public entry point is a trial and the real product is sales-led.
5. Ahrefs Brand Radar. A convenient add-on if Ahrefs is already your SEO base. Our catalog notes its prompts are modeled from keyword data and its chatbot indexes refresh roughly monthly.
Who should pick which
If you only want a yes or no this week, use the free Promptwatch Explore plan and a spreadsheet. If you want to know why you are absent and act on it, you need citation and crawler data too, which is where an end-to-end platform pulls ahead of pure trackers. If you are an enterprise team with procurement requirements, shortlist Profound or Scrunch AI Enterprise alongside Promptwatch Business.
What the answer can and cannot tell you
A tool can tell you whether ChatGPT mentions you, how often, next to whom, and from which sources. It cannot make ChatGPT recommend you. Expect variation between runs, and judge progress by trends across a panel of prompts, not single answers.