This tool asks its questions in Spanish and covers Mexican-market categories, so the topics come back in Spanish. Everything around them is in English.
Free tool
Measuring how AI sees you starts with knowing what to ask it.
When somebody asks ChatGPT what to buy, the model answers with whatever topics from your category it has to hand. A category is the kind of product you compete in, not your brand: craft beer, instant coffee, laundry detergent. This tool builds the list of topics that get discussed there.
Those topics apply to any brand in the category, not only yours, because they describe what people discuss when they discuss the product rather than who makes it. They are where the questions you put to the models come from, and that is what makes them useful: the more topics your questions cover, the fewer blind spots your measurement leaves. A topic you never asked about is a topic where you do not know whether you appear.
What happens when you ask
The questions go to an AI model, the same kind of model that answers the person deciding what to buy. How you ask changes what it answers, and the size of that change is the point of this page.
The category
craft beer
The category you compete in. Never your brand name.First, a single question
5 topicsWhat do craft beer buyers take into account when choosing which one to buy?
- taste
- price
- brand
- alcohol content
- packaging
The question anyone would ask, with no extra instructions. The model answers with whatever it has to hand.Then, one question per dimension
~90 topicsFor each dimension of the category, give me topics that shape the choice. At least one a brand manager would not raise on their own.
- beer for the Sunday barbecueOccasion
- whether it survives the trip in the coolerPracticality
- hops too bitter for a newcomerSensory
- 473 ml can against a 355 bottleFormat
- breweries calling themselves craft while owned by a giantTrust
- gluten-free options for coeliacsHealth
and 84 more topics
A dimension is one angle you can look at the category from. There are fourteen, and each one uncovers topics the plain question never reaches.Finally, the repeats get merged
Two topics saying the same thing in different words count once. Only when there is no doubt, because losing a topic costs more than keeping two similar ones.We will send you the full list
It arrives as a CSV built as a review sheet: a column to mark which topics matter, another to fix the wording, and room at the end to add the ones missing. No account and no password.
These figures come from one example exploration. Every category returns its own number, and the one you will see below is yours.