COPILOT IN POWER BI

Power BI Copilot Prompts: DAX, Dashboards, And Data Exploration Examples

Copilot in Power BI is a chat-based assistant that can draft DAX measures, build report pages, and answer questions about your data. Power BI Copilot prompts work best when they are specific: naming your actual tables, columns, and business terms instead of asking generic questions.

Key Takeaways

  • Copilot in Power BI requires a paid Fabric capacity (F2 or higher) or Power BI Premium (P1 or higher). Trial capacities and free SKUs are not supported.
  • Prompts can run up to 10,000 characters across every Copilot surface in Power BI, so there is room to give real context, not just a one-line question.
  • Model owners need to prepare their semantic model for AI first. Without that prep, Copilot can misinterpret your data and return generic or misleading answers.

Who This Is For

Power BI report authors and BI analysts who already build reports and want a practical library of Copilot prompts for DAX, report creation, and data exploration.

In This Guide

What Copilot Can Actually Do In Power BI

Microsoft splits Copilot in Power BI into two audiences, and the prompts you write should match which one you are.

  • For report authors: Copilot can create and edit report pages, summarize the underlying semantic model, add a narrative summary visual, write DAX queries, and draft descriptions for semantic model measures.
  • For business users: Copilot can answer questions about an open report through the Copilot pane, or search and analyze any report, semantic model, or Fabric data agent through the standalone Copilot experience.

Both roles hit the same wall: Copilot answers from your semantic model when a question relates to your data, and from the underlying language model’s general knowledge otherwise. If your data isn’t modeled cleanly, or your business terms aren’t documented, Copilot has nothing reliable to draw from.

Flow diagram showing a Power BI Copilot prompt moving from a prepared semantic model through the Copilot engine to a DAX measure, report visual, or narrative summary, then human review and validation before publishing

Figure 1: The path from a Power BI Copilot prompt to a published, human-validated report.

Prompt Examples For DAX Measures

DAX generation is one of the most requested Copilot use cases, and it is also where vague prompts fail the most obviously. Reference your real measure and column names every time.

Goal Prompt
Basic aggregation measure “Write a DAX measure named Total Rental Revenue that sums the RentalAmount column from the FACT_RENTALS table.”
Time intelligence measure “Write a DAX measure that calculates year-over-year growth in Total Rental Revenue, using the DIM_DATE table marked as a date table.”
Filtered measure “Write a DAX measure for Average Rental Days that only includes rentals where VehicleClass equals ‘SUV’, using CALCULATE and a filter on DIM_VEHICLE.”
Measure explanation “Explain what this DAX measure does in plain language, step by step: [paste your measure].”
Measure description “Draft a one-sentence description for the measure Total Rental Revenue, suitable for a business user browsing the semantic model.”

Pro tip: Ask Copilot to explain a DAX measure it just wrote before you use it. Explaining forces it to restate the logic in plain language, which is often where you catch a filter context mistake.

Prompt Examples For Report And Dashboard Creation

Copilot’s report-building capability works from a written prompt describing what you want to see, not from a blank canvas.

Goal Prompt
Create a report page “Create a report page showing Total Rental Revenue by month, Vehicle Class, and Rental Location, using a combination of a line chart and a bar chart.”
Add a KPI summary “Add a card visual showing Total Rental Revenue for the current quarter compared to the prior quarter.”
Generate a narrative visual “Add a narrative visual that summarizes the key trends in Total Rental Revenue by Vehicle Class for this report page.”
Summarize an existing report “Summarize the key insights on this report page in three bullet points a regional manager could read in ten seconds.”
Redesign for an audience “Suggest a layout for this report page aimed at store managers who need to check daily rental volume in under 30 seconds.”

Prompt Examples For Data Exploration

These prompts work through the Copilot pane on an open report or the standalone Copilot experience, and they lean on Copilot’s ability to answer questions using your semantic model directly.

Goal Prompt
Ad hoc question “Which rental location had the highest Total Rental Revenue last month?”
Trend question “Has Average Rental Days increased or decreased over the last four quarters?”
Comparison question “Compare Total Rental Revenue between SUV and Sedan vehicle classes for this year.”
Anomaly question “Are there any rental locations where revenue dropped more than 15 percent compared to the prior month?”
Find content “Find any reports or semantic models related to vehicle rental revenue that I have access to.”

Copilot Limitations You Should Know

  • Paid capacity is required. Copilot needs a paid Fabric capacity (F2 or higher) or Power BI Premium (P1 or higher). A Power BI Pro or Premium Per User license alone is not enough, and trial capacities are not supported.
  • Prompts are capped at 10,000 characters. That is generous, but a prompt that pastes an entire data dictionary will still hit the ceiling.
  • Data prep changes the output quality. Without preparing a semantic model for AI, Copilot can struggle to interpret data correctly and produce generic or misleading responses.
  • Some experiences are still in preview. The Copilot pane for reports is generally available, but the standalone Copilot experience and Copilot in apps were in preview at the time Microsoft published its current documentation.
  • Responses can be cached. The same exact prompt against an unchanged semantic model within a rolling 24-hour window returns a cached response rather than a fresh one.
  • Sovereign clouds and some regions are not supported. Standalone Copilot and Copilot in apps are unavailable in a list of regions Microsoft maintains, and sovereign clouds are not supported at all.

Prompt-Writing Checklist

Use this before you send a Copilot prompt in Power BI, especially for DAX or report-building tasks.

  • The prompt names real table and column names, not generic placeholders
  • The prompt states the business term you want (for example, “Total Rental Revenue,” not just “revenue”)
  • For DAX, the prompt specifies the filter conditions and table relationships involved
  • For report prompts, the prompt names the visual type or explicitly asks Copilot to choose one
  • You have asked Copilot to explain its own output before using it
  • You have confirmed your organization’s Fabric capacity and admin settings support the Copilot experience you’re using

Frequently Asked Questions

Can Copilot in Power BI write DAX measures for me?

Yes. Report authors can ask Copilot to write DAX queries and measures directly, and can also ask it to draft plain-language descriptions of existing measures. Always review the generated DAX against your actual filter context before publishing it.

Yes. Copilot requires a paid Fabric capacity (F2 or higher) or Power BI Premium (P1 or higher). A standalone Power BI Pro or Premium Per User license is not sufficient on its own.

This usually means the semantic model has not been prepared for AI. Without clear business terms, descriptions, and clean relationships, Copilot has to guess at context, which leads to generic or misleading responses.

Not entirely. The Copilot pane for reports is generally available, but the standalone Copilot experience and Copilot in apps are unavailable in several regions and are not supported in sovereign clouds.

Write Prompts That Match What Copilot Can Actually Do

Power BI Copilot prompts work best when they read like instructions to a new analyst who knows nothing about your business yet: specific table names, specific business terms, and a specific ask. The Copilot in Microsoft Fabric & Power BI course covers preparing your semantic model and building AI-assisted reports end to end, and connects to Chapters 9 and 11 of Business Intelligence in the Age of AI.

Related Free BI Resources

Not ready to enroll yet? Start with these free BI Academy resources.

Keep your BI skills sharp – follow for more insights!

Leave A Comment