Freeze the setup
Record the ChatGPT surface, model label, web-search setting, date, locale, and whether the session is fresh.
Free interactive test
ChatGPT brand visibility test
A ten-question manual test for checking whether ChatGPT discovers, describes, compares, and cites your brand under consistent conditions.
Published:
10-prompt manual test
Record what ChatGPT actually returns
Replace the bracketed fields, use fresh sessions with consistent web search settings, save each answer, and classify the visible result.
0/10
Tested
0%
Mention rate
0%
Citation rate
- 01What are the best [category] options for [target audience]?
- 02Which [category] providers should a [buyer role] evaluate?
- 03What tools help [target audience] solve [problem]?
- 04Which [category] is best for [use case]?
- 05What are the best alternatives to [competitor]?
- 06Which [category] vendors publish the clearest methodology?
- 07What evidence should I ask for before buying [category]?
- 08What is [brand], and what does it do?
- 09What are the strengths and limitations of [brand]?
- 10How does [brand] compare with [competitor]?
Save the evidence
Keep the full answer, visible source URLs, prompt wording, and any errors instead of recording only a yes or no.
Repeat carefully
Use the same prompts and conditions for a rerun. Label a changed prompt set as a new benchmark.
Common questions
Use the resource without overstating the result
How do I test whether my brand appears in ChatGPT?
Define buyer questions, replace the placeholders, use consistent fresh-session and web-search conditions, save each answer, and record whether the brand appears and whether the claimed domain is cited.
Why are only three prompts brand-named?
Neutral prompts test whether ChatGPT discovers the brand independently. Brand-named prompts instead test recognition, accuracy, and competitive framing and should be reported separately.
Will ChatGPT return the same answer every time?
No. Answers can vary by model, product surface, session context, search behavior, location, and time. Record the conditions and treat a small manual test as directional.
Does a mention prove that ChatGPT recommends the brand?
No. A mention may be incidental, comparative, negative, or prompted by the brand name. Record prominence, recommendation language, sentiment, accuracy, and citations separately.
Evidence before conclusions
See a complete benchmark before running your own
Review the questions, stored answers, citations, competitors, coverage, and five prioritized actions in the sample report.