Free tool
Find out if assistants name you
This generates the prompts you should run yourself in ChatGPT, Claude, Gemini, and Perplexity to find out whether assistants name your brand, plus a sheet for scoring what comes back. It calls no AI model. It gives you a repeatable test protocol, and running it is your job.
Last reviewed 27 August 2026
Result
Checks in this pack
144
12 prompts across 4 assistants, 3 runs each, roughly 3.6 hours of work
Prompt pack
AI visibility test pack for Acme Analytics Run each prompt 3 times in each of: ChatGPT, Claude, Gemini, Perplexity. Use a fresh chat with no memory or personalisation for every run. 1. What are the best product analytics options available right now? 2. I am a founder and I need to work out which accounts are actually using the product. What should I use and why? 3. Who are the main companies in product analytics? 4. What does Acme Analytics do, and who is it for? 5. Is Acme Analytics any good? What do people say about it? 6. What are the alternatives to Mixpanel? 7. Compare Acme Analytics and Mixpanel for a founder. 8. Which fits a small team better, Mixpanel or Amplitude? 9. What should a founder look for when choosing product analytics? 10. Which product analytics option is best value for a team under twenty people? 11. What are the common complaints about product analytics products? 12. Recommend three product analytics options and explain the trade-offs between them.
Scoring sheet
| # | Prompt | ChatGPT | Claude | Gemini | Perplexity |
|---|---|---|---|---|---|
| 1 | What are the best product analytics options avail... | ___ | ___ | ___ | ___ |
| 2 | I am a founder and I need to work out which accou... | ___ | ___ | ___ | ___ |
| 3 | Who are the main companies in product analytics? | ___ | ___ | ___ | ___ |
| 4 | What does Acme Analytics do, and who is it for? | ___ | ___ | ___ | ___ |
| 5 | Is Acme Analytics any good? What do people say ab... | ___ | ___ | ___ | ___ |
| 6 | What are the alternatives to Mixpanel? | ___ | ___ | ___ | ___ |
| 7 | Compare Acme Analytics and Mixpanel for a founder. | ___ | ___ | ___ | ___ |
| 8 | Which fits a small team better, Mixpanel or Ampli... | ___ | ___ | ___ | ___ |
| 9 | What should a founder look for when choosing prod... | ___ | ___ | ___ | ___ |
| 10 | Which product analytics option is best value for... | ___ | ___ | ___ | ___ |
| 11 | What are the common complaints about product anal... | ___ | ___ | ___ | ___ |
| 12 | Recommend three product analytics options and exp... | ___ | ___ | ___ | ___ |
How to score each answer
- 2 points: your brand is named in the first three options, or named directly in the answer to a direct question.
- 1 point: your brand appears somewhere lower down, in a longer list, or only after a follow-up question.
- 0 points: your brand is absent, or the assistant describes you incorrectly.
- Record which sources the answer cites where the assistant shows them. Those pages are your actual competition for the citation, and they are usually review sites, comparison posts, and forum threads rather than your own site.
- Repeat the whole pack monthly. A single run is a snapshot of a system that changes weekly, and the trend is the only part worth acting on.
This tool deliberately calls no AI model. Any tool that queries assistants for you is running one prompt, once, through an API that behaves differently from the consumer product your buyers actually use, and then reporting it as your visibility score. Running the prompts yourself takes an afternoon and produces something you can trust, including the part that matters most, which is reading what the assistant says about you when it does name you.
How this works
What the numbers mean.
- 01The prompts are generated across four intents: open category questions where you might be listed, direct brand questions that test whether the model knows you at all, comparison questions against named competitors, and problem-led questions phrased the way a buyer would type them. Visibility differs sharply between the four, and averaging them hides that.
- 02Each prompt is run several times per assistant because answers are not deterministic. Being named once in three runs and named three times in three runs are very different positions, and one run cannot tell them apart.
- 03The scoring sheet separates being named early from being named at all, because placement in a list drives selection in the same way it does in search results.
Assumptions and limits
- This is a manual protocol. It produces no automated score, and it is honest about needing an hour or two of your time.
- Use a fresh chat with memory and personalisation switched off. Any assistant that remembers you have discussed your own company will name it, which invalidates the whole test.
- Results vary by account, region, model version, and week. Treat any single run as a data point, never as a measurement.
Questions about this tool
Why does AI visibility matter now?
Because a growing share of search ends without a click. SparkToro's clickstream analysis found 68.01 percent of Google searches ended without a click, and Pew Research found people clicked a traditional result in 8 percent of visits where an AI summary appeared, against 15 percent where none did. If the answer is assembled before anyone reaches a website, being named inside that answer is the visibility.
Why not just use a tool that checks this automatically?
Most of them send one prompt through an API and report the result as a score. API responses differ from the consumer apps your buyers use, personalisation and memory are absent, and one run of a non-deterministic system is noise. The manual version costs an afternoon and gives you the qualitative half as well, which is how the assistant describes you when it does mention you.
What do I do when my brand does not appear?
Look at what the assistant cited instead. Answers are usually assembled from comparison articles, review platforms, community threads, and documentation rather than from company home pages. Getting accurately represented on those sources moves the needle more than anything you publish on your own domain, which is an uncomfortable but consistent finding.
How often should I run the pack?
Monthly is enough for most companies, quarterly if the category moves slowly. Model versions change, indexes refresh, and the sources being cited shift, so a single measurement ages quickly. Keep the sheet and watch the trend, because the direction of travel is far more actionable than any individual answer.
This tool is free and there is nothing to sign up for. If you would rather have the work done than calculate it, that is what Zway does.