Be Specific About What You Want
Be Specific About What You Want
Clear questions produce clearer results. Broad prompts like “Analyze this data” leave too much open to interpretation. Compare that with “Show the top 10 customers by revenue, including average order value and purchase frequency.”Specificity applies to scope and time frames as well. “Sales by region” is ambiguous, while “Sales by region for Q3 2024 compared to Q3 2023” defines exactly what should be analyzed.You don’t need to know the outcome in advance—exploration is encouraged—but narrowing the question helps the system focus on the right dimensions.
Reference Files Where Needed
Reference Files Where Needed
When multiple files are available, referencing them by name can help disambiguate your request. For example, “According to the Q3 Report…” or “Compare the projections in the Budget spreadsheet to actual results.”If no ambiguity exists, implicit references work just as well.
Build on Previous Responses
Build on Previous Responses
Follow-up questions are often more effective than trying to specify everything upfront. Start with an initial query, then narrow or adjust based on the output.This incremental approach mirrors how most real analysis is done.
Ask for Explanations When Needed
Ask for Explanations When Needed
If a result isn’t clear, ask how it was derived. Questions like:
- “How was this calculated?”
- “Why were these records excluded?”
- “Explain this metric in simpler terms”
Request Specific Formats
Request Specific Formats
You can guide how results are presented by asking for a specific format:
- “Summarize this in three bullet points”
- “Provide a one-paragraph executive summary”
- “Show this as a comparison table”