Why our AI asks "which metric?" before drawing a chart

A small design decision we keep coming back to: when a request is ambiguous, PowerDataChat asks a clarifying question instead of guessing.

Example: "Chart revenue and margin by month." Revenue is in currency, margin is a percentage. Put both on one chart and it becomes hard to read. So instead of picking for you, it asks which metric you want, or how you want them shown.

Three related behaviours follow the same idea:

• Starter questions. An empty chat box is intimidating. After a file is connected, the AI suggests questions based on that file, so a new user has somewhere to start.
• Honest "missing data" answers. If you ask about a column that does not exist, it says what is missing and suggests alternatives instead of inventing values.
• Follow-ups keep context. "Now only for 2026" works without restating the whole question.

None of this is clever AI. It is the behaviour you expect from a good human analyst: ask when unclear, say when the data cannot answer, remember what you were talking about.

What do your users do when your product's input is ambiguous? We would like to hear how others handle it.