The exact prompts to test yourself.
Enter your category and we generate the prompts your buyers actually ask AI, across discovery, comparison, trust and pricing intent, with a scoring sheet to track results month over month.
The audit that actually works.
Describe your category
Enter what you sell in the words a buyer would use, plus your brand, competitors and market if relevant.
Run the prompts yourself
Paste each into ChatGPT, Gemini and Perplexity using a fresh session. Record whether you are named, where, and who appears instead.
Repeat monthly
One snapshot tells you little. A monthly record shows whether your visibility is improving, which is the only measure that matters.
The manual audit beats most dashboards.
This is a deliberately low-tech tool, and that is the point. Industry analysts covering AI visibility keep arriving at the same recommendation: query your own brand and key topics across the major engines, note what gets cited, and note where competitors appear and you do not. That manual audit reveals more about your current position than most dashboards do.
The reason is that the AI visibility tracking market is young and expensive. Enterprise platforms sample prompts on your behalf and report aggregate scores, which is genuinely useful at scale, but you are trusting their prompt selection. Nobody knows your buyers' actual language better than you do.
What this tool gives you is structure. Thirty-plus prompts spanning discovery, comparison, trust, pricing and local intent, so you test the full journey rather than only the flattering question of whether your brand name is recognised.
Run it once and you have a baseline. Run it monthly and you have a trend, which is what tells you whether your AEO work is actually moving anything. Set expectations properly though: answers vary by user, phrasing and region, so look for patterns across several runs rather than reacting to any single result.
Prompt testing questions.
Why test manually instead of using a tool?
Manual testing gives you the real rendered answer your buyer would see, in your own words rather than a vendor's prompt set. Paid platforms are valuable at scale and worth it for large brands, but a structured manual audit costs nothing and is often more revealing for a single business.
How often should I run these?
Monthly is a sensible cadence for most businesses. AI answers shift as models update and as competitors publish, so quarterly is too slow to catch movement and weekly produces noise rather than signal.
Should I use a fresh chat each time?
Yes. Prior conversation influences responses, and personalisation can skew results toward what you have asked before. Use a new chat, and ideally a logged-out or incognito session, so you see something closer to what a stranger would get.
What should I record?
Whether your brand is named at all, its position in the answer, which competitors appear, what the engine says about you, and the date. The competitor data is often the most useful part, because it shows you which sources the engine trusts in your category.
What do I do with poor results?
Look at what the engines cite instead of you. Usually it is a comparison site, a Reddit thread or a competitor's well-structured page. That tells you exactly where the authority gap is, and it is far more actionable than a score.
Poor results? We fix that.
Send us your prompt results and we will tell you why the engines are naming someone else, and exactly what it takes to change that.