Collect a brief to analyze competitors or topics for work. Return JSON {"question":"...","targets":"...","dimensions":"...","output_format":"..."}. question - what and why we analyze; targets - specific objects of analysis (competitors, products, approaches); dimensions - 4-6 comparison dimensions according to the same standards (price, assortment, service, strengths/weaknesses); output_format - result form for the meeting: a table (or structured conclusions) is required and a section “what should we do from this” with 2-3 actions.
Mechanics: analysis of competitors without structure produces scattered notes that will not come together at the meeting into a picture. The value of a working analysis is in comparability: all objects are run through the same measurements, so you can see who is strong where. Levers. Concrete objects: name them, otherwise the model will take abstract ones. Unified dimensions: 4-6 axes applied to all. Rely on facts: ask for observable indicators (prices, hours of operation, reviews), not guesswork. Inside trick: add the dimension “how the client gets / how he wins” - not a dry parameter, but a hypothesis about the reason for success. This is what analysis is done for, but beginners forget to ask. Second technique: end with the conclusion “what should we do from this” - 2-3 specific actions. A debriefing without a “so what now” in a meeting is useless; make the model come up with recommendations. A typical mistake: collecting facts and not drawing a conclusion. Data without interpretation is homework, not research. Always demand “that’s why we should.”
Unlock access to submit solutions for instant AI review.