Reference: Decision Intelligence Prompts

You can use these prompt examples to help you explore data and model scenarios.

Use Case	Sample Prompt
Ad hoc analysis	Walk me through the biggest movers in last quarter's P&L and draft a one-paragraph narrative for the board readout.
Cohort and Pareto analysis	Which 20% of cost centers account for 80% of the budget variance? Show me the Pareto curve.
Contribution analysis	What are the top 5 drivers of the margin decline? Break it down by region and product line.
Driver-based reforecasting	Using this pipeline data and the headcount plan, build a Q4 revenue reforecast. Show what happens if we slip the new product launch by one quarter.
Monte Carlo simulation	Run a Monte Carlo simulation on Q4 revenue using the historical variance in our pipeline conversion rates. Show me the P10, P50, and P90 outcomes.
Regression and trend analysis	Fit a regression to our quarterly revenue over the last 3 years and project the next 4 quarters. What's the R-squared and which drivers are most significant?
Root cause analysis	What's causing the spike in SG&A? Decompose by cost center and show me the biggest movers.
Scenario modeling	Build three scenarios from this baseline: optimistic, conservative, and a stretch case. Compare them side by side.
Sensitivity analysis	Show me how net revenue changes if we vary the conversion rate from 10% to 25% in 5-point increments.
Variance investigation	Why is EMEA revenue down in Q3?
Vendor analysis	Flag vendors with year-over-year increases above 15% and model the impact of consolidating the top 5.
Workforce planning	Model the cost of a 4% merit increase across the organization plus 8 new hires in Q2. Show the fully loaded cost projection.

Prompt Best Practices

Upload all relevant data at the start to provide DI with the full picture.

Example: Upload actuals alongside budget and headcount before you start asking questions.

Frame the decision, not just the data. Example: Use the prompt Help me decide whether to delay the product launch
rather than Show me the revenue data
.

Push back when something looks wrong. Example: To see a step-by-step explanation of how DI provided the incorrect information, enter into the chat: That doesn't look right. Walk me through how you calculated that.

Click the thumbs up or thumbs down icon to provide feedback on the quality of a response. If you click thumbs down, you can provide feedback on how the response can be improved.

Review before acting. DI is a decision-support tool. Always review outputs before submitting changes to plan.