Use Power BI Copilot to Query Demand Data in Plain English

What This Does

Power BI Copilot lets you ask questions about your demand planning data in plain English and get charts, summaries, and reports back — without writing DAX formulas or waiting for a BI analyst. For demand planners who need custom data views for ad hoc management questions, this eliminates the BI request backlog.

Before You Start

• Power BI Pro or Premium license (Copilot features require these tiers)
• Access to a Power BI report or dataset connected to your demand/sales data
• Logged in to Power BI service (app.powerbi.com)

Steps

1. Open Your Demand Planning Report or Dataset

In Power BI service, open the report or workspace containing your demand data. Look for a report that has actuals, forecasts, inventory, or order data.

What you should see: A report with existing visualizations, or a blank canvas connected to your data model.

2. Open the Copilot Pane

Click the Copilot button in the report toolbar (sparkle icon in the top menu bar). A Copilot panel opens on the right side.

What you should see: A chat interface where you can ask questions about the data in the current report.

3. Ask Your First Data Question

Type a plain English question:

Show me forecast accuracy (MAPE) by product category for the last 3 months, sorted from worst to best accuracy.

What you should see: Power BI Copilot either generates a new visual directly or points you to the existing visual that answers the question. Click "Add to report" if it creates something new you want to keep.

4. Use It for Ad Hoc Management Questions

When your VP asks "which SKUs are driving the most inventory risk right now?" — instead of building a new report, type:

Which SKUs have the highest days of supply vs. the demand plan? Show top 10 sorted by excess days.

What you should see: A table or visual showing your highest-risk inventory positions — the answer to a question that would have taken 30 minutes to build manually.

5. Generate a Narrative Summary

Power BI Copilot can also write text summaries of your data:

Write a 3-sentence summary of this month's demand performance vs. plan. Highlight the biggest category variance and the top product driving it.

What you should see: A text block you can copy directly into your S&OP presentation notes.

Real Example

Scenario: You're 2 hours before the monthly demand review and your VP asks: "Can you quickly show me which regions underperformed the forecast by more than 10% this month?"

What you do: Open Power BI → Copilot pane → type: "Show me a bar chart of actuals vs. forecast variance by region for this month. Flag any region where actuals were more than 10% below forecast."

What you get: A visual ready in 60 seconds. You copy it into a PowerPoint slide, add a one-sentence commentary (which Claude drafts for you), and walk into the meeting prepared.

Tips

• Copilot works best when your data model has clear, descriptive field names — "Region," "Category," "MAPE," "ActualUnits" are easier than "Rgn_cd," "Cat_id," "accuracy_pct"
• Ask follow-up questions in the same session — Copilot maintains context within a conversation, so "now filter this to just the beverage category" works as a follow-up
• If Copilot generates a visual that's close but not quite right, describe the change: "Make it a line chart instead" or "Add a target line at 15% MAPE"

Tool interfaces change — if a button has moved, look for similar AI/magic/smart options in the same menu area.