LearnAI
Новинки AI · Данные и таблицы · Lesson 3/10
← To the catalog
Task

Ask a question, don't "look"

Formulate one precise question about the data, tied to the solution. Return JSON {"data":"...","question":"...","decision":"..."}: data — context; question - a specific measurable question (with a cut-off, metric and period), to which there is an unambiguous answer from the table; decision - what decision will you make depending on the answer.

Unlock access to submit solutions for instant AI review.
Ask for case information
💡 A piece of theory

A vague question produces a vague answer. “How are sales?” - and the AI ​​throws out ten disparate facts, none of which help solve the problem. A good question is measurable: it has a slice (breakdown - for example, by time of day), a metric (what we count - revenue or the number of cups) and a period (per month). Three pillars - and the answer becomes clear. Insider: tie the question to the solution in advance. “If evening revenue is less than 20% of daytime revenue, I reduce my hours.” This turns data into action and cuts out interesting but useless facts: you only ask what will change your step. Second trick: ask for comparisons, not absolutes. “How much in the evening COMPARED to the day” is more informative than a single number - 40 cups is a lot or a little, it is only clear next to another number. A typical mistake: ask a general question “look at sales” and expect the AI ​​to guess what decision you are making. He is not a telepath - decide for yourself, then ask.

How is it assessed · passing 70

  • 1Valid JSON without text around it25%
  • 2There are fields data, question, decision25%
  • 3The question is measurable and related to the solution50%
Your promptClaude ⌄
🔒

Unlock access to submit solutions for instant AI review.