Asking ChatGPT "do you know my company?" once tells you very little. Answers change between runs, between users and depending on whether the assistant searched the web. A useful check is a small, repeatable sample: the right prompts, asked the same way, recorded over time. Here's how to do it by hand in about 30 minutes, and which tools can do it for you.
Step 1: Write the prompts your buyers use
Leave your brand name out of discovery prompts. You want to know whether the assistant brings you up on its own. Aim for 10 to 20 prompts across these types:
- Category: "What are the best [category] tools for [audience]?"
- Alternatives: "What are good alternatives to [biggest competitor]?"
- Problem: "How do I [job your product does]?"
- Comparison: "[Your brand] vs [competitor]: which is better for [use case]?"
- Brand facts: "What is [your brand] and how much does it cost?" This one checks whether what the assistant says about you is accurate.
Step 2: Remove personalization
Your own ChatGPT account knows what you work on, which skews the answer. In ChatGPT, start a Temporary Chat and choose Unpersonalized before your first message: OpenAI says that mode doesn't use memory or custom instructions, and temporary chats stay out of your history. A logged-out browser window works too. For other assistants, use a private window or an account that has never discussed your company.
Step 3: Note whether the assistant searched
ChatGPT can answer from what the model learned in training or by searching the web, and when it searches it shows sources. Note which one happened. If you only appear in answers with search, your pages and the sites that mention you are working, but the model itself doesn't know you well. If you're missing even with search, look at the sources it cited instead of you: those are the pages to get listed on.
Step 4: Repeat and record
Run each prompt at least two or three times, because the same prompt can produce a different list each time. Record the results in a sheet so you can compare month to month:
date,assistant,prompt,searched,mentioned,position,competitors_named,sources_cited 2026-10-14,ChatGPT,"best uptime monitoring for startups",yes,no,,"Acme, Beta","g2.com, reddit.com"
Step 5: Check the other assistants
Buyers don't all use ChatGPT. Repeat the core prompts in Perplexity, which always shows its sources, Claude, Gemini, Google's AI Mode and Microsoft Copilot. Each uses different models, indexes and crawlers, so a gap in one and not another usually points to a crawler or indexing problem rather than your content.
What to do with the results
- Not mentioned, competitors are: study the cited sources and get onto them (review sites, comparison articles, community threads). Make sure your own comparison and alternatives pages exist and are crawlable.
- Mentioned, but with wrong facts: fix the source of truth. Put pricing, features and positioning in plain HTML, add Organization and product schema and an llms.txt, and keep descriptions consistent across profiles.
- Not mentioned anywhere, even for brand prompts: check that AI search bots can reach you with the AI crawler checker, then work through how to get recommended by ChatGPT.
Tools that automate this
RankRoot's free AI mention check asks Google Gemini, through the official Gemini API, a set of buyer prompts built from what you sell and shows whether you're named, where you appear in the list, which competitors are recommended instead and which sources are cited. It covers Gemini only, not ChatGPT, and the free version runs one check of 3 prompts a day.
Paid trackers query several assistants on a schedule and chart your share of voice. Options include Profound, Peec AI, Otterly.ai, the Semrush AI Visibility Toolkit and Ahrefs Brand Radar. Check each vendor's current coverage and pricing before buying, and keep in mind that answers collected through an API can differ from what people see in the consumer apps.