To know what AI says about your brand, ask each major assistant (ChatGPT, Claude, Gemini, Perplexity) the questions your buyers actually ask: recommendation prompts, comparisons, and direct brand questions, then record whether and how you appear. One chat is not an answer, it is a single sample, so the method matters more than any individual response.
Below is the manual method, the blind spots it cannot escape, and what systematic scanning adds.
The manual method: ask five prompt types
Open a fresh session in each assistant and work through these prompt types, replacing the placeholders with your category and brand:
- Recommendation: "what is the best [category] for [buyer situation]", the highest-intent prompt type
- Comparison: "[your brand] vs [competitor]", reveals how AI frames your strengths and weaknesses
- Alternatives: "alternatives to [market leader]", where challenger brands most often surface
- Direct brand: "what is [your brand] and what does it do", tests whether AI knows you at all and gets the facts right
- Pricing and fit: "is [your brand] good for [use case]", surfaces the caveats AI attaches to you
Where manual checking goes blind
The manual method has structural blind spots. AI answers vary between runs, so the response you saw may not be the one most buyers get. Your own account's memory and chat history can skew results toward brands you already searched. And you can only sample a handful of prompts by hand, while buyers phrase the same need dozens of ways.
The deeper problem: AI providers do not publish prompt logs. There is no Search Console equivalent for ChatGPT or Perplexity, a gap explained in why no Search Console exists for AI prompts. You cannot look up what people asked; you can only sample what AI answers.
Systematic scanning: sampling done properly
A systematic scan runs a fixed panel of buyer prompts across multiple providers under clean conditions, then scores the results: were you mentioned, at what position, with what sentiment, and against which competitors. Repeating the same panel over time turns anecdotes into a trend line.
This is what the free brand scan automates: it tests your brand across ChatGPT, Claude, Gemini, Perplexity, and Google AI Overview using unbranded buyer prompts, and returns a GEO score with a per-provider breakdown. If you find you are being outnamed, the next step is diagnosing why AI recommends competitors instead of you.
See what AI actually says about you
Run the free scan and get a provider-by-provider readout of how ChatGPT, Claude, Gemini, and Perplexity describe your brand on real buyer prompts.
Frequently asked questions
How do I find out if ChatGPT mentions my company?
Ask ChatGPT unbranded buyer questions for your category in a fresh session, such as "best [category] for [use case]", and note whether your company appears. Repeat across several phrasings, because answers vary between runs. For a reliable read, run a fixed prompt panel on a schedule rather than one-off checks.
How do I check if ChatGPT knows about my business?
Ask it directly: "what is [business name] and what does it do". If it answers accurately, the model recognizes your entity. If it hedges, confuses you with another company, or invents details, you have an entity problem to fix before chasing recommendations.
What tools show how AI describes my company?
AI visibility tools sample many prompts across providers and report mentions, position, and sentiment. YouGotRanked's free scan covers ChatGPT, Claude, Gemini, Perplexity, and Google AI Overview; platforms like Profound, Peec, and Otterly also track AI answers on subscription models.
Why do I get different answers each time I ask?
AI responses are probabilistic, so the same prompt can produce different brand lists on different runs. Model version, search mode, and account memory add more variance. That is why single checks mislead and sampling across repeated runs is the only honest measurement.
Related answers