How to Use Copilot in Power BI Desktop
Microsoft's Copilot for Power BI is changing the game by automating tedious report building and complex calculations, letting you focus on what the data actually means. This guide shows you how to get it running and start using it today to create reports, write technical queries, and even generate data summaries with simple, conversational language. We’ll walk through the setup process and show practical examples for a faster, smarter workflow.
What is Copilot for Power BI?
Copilot for Power BI is an AI assistant built directly into Power BI Desktop and the Power BI service. It functions like a data-savvy partner you can talk to. Instead of manually clicking through menus, dragging fields, and writing complex code, you simply tell Copilot what you need in plain English. It understands your data model and can take action for you.
Its capabilities currently fit into three main areas:
- Automated Report Creation: Ask it to build an entire report page based on an analytical goal, like "create a page showing a summary of our marketing campaign performance." Copilot will select the right data and build appropriate charts for you.
- DAX Generation and Explanation: It helps you write and understand Data Analysis Expressions (DAX), the formula language for Power BI. You can ask for a calculation like "YoY GMV % Change" and Copilot will generate the DAX code, or ask it to explain a complex formula you found in an existing report.
- Narrative Summaries: Copilot can generate a text summary of a visual or an entire report page, pulling out key insights and trends in paragraph or bullet-point form. This is incredibly useful for adding context to your dashboards.
Getting Started: How to Enable Copilot in Power BI
Before you can start chatting with your data, you need to make sure Copilot is enabled. This involves checking a few administrator settings and flipping a switch in Power BI Desktop. If you're not an admin, you may need to ask your IT or BI team for help with the first two steps.
1. Check Your Licensing and Capacity
Copilot is considered a premium feature. To use it, your organization needs to have either a Power BI Premium (P1 SKU or higher) or Fabric (F64 SKU or higher) capacity. The workspace you publish your report to must be assigned to one of these capacities. You don’t need a specific per-user license beyond a standard Pro or Premium Per User (PPU) license, but the backend infrastructure has to meet these requirements.
Free PDF · the crash course
AI Agents for Marketing Crash Course
Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.
2. Enable Copilot in Your Fabric Admin Settings
A Fabric or Power BI administrator needs to turn on Copilot at the tenant level. Once enabled, they can decide to make it available to the entire organization or just specific security groups.
Here’s the path for admins:
- Log into your Fabric/Power BI admin portal.
- Go to Tenant settings and scroll down to the "Copilot and Azure OpenAI Service settings" section.
- Find the setting titled "Copilot," open it, and click Enabled.
- Apply it to the entire organization or specify security groups for access.
There is also a second setting on that page that sends data to Azure OpenAI, which must also be enabled.
3. Enable Copilot as a Preview Feature
Once the admin has enabled it, you need to activate it in your own Power BI Desktop application. This is because Copilot is still technically a "preview feature."
- Open Power BI Desktop.
- Go to File > Options and settings > Options.
- Navigate to the Preview features section.
- Check the box for Copilot.
- Click OK and restart Power BI Desktop.
After restarting, you should see a new "Copilot" button on the ribbon in the Report view. You're now ready to use it.
How to Build Reports Instantly with Copilot
Creating report pages is one of Copilot's most impressive time-saving features. It’s perfect for quickly scaffolding a new report or exploring a dataset without knowing which visuals to build first.
In the Report view, click the Copilot button in the ribbon to open the Copilot pane. This is your chat window.
Starting with a Broad Prompt
Your first prompt can describe a general goal. Think about what business question you want to answer. Power BI will scan your data model to find relevant fields to answer that question.
For example, using a simple sales dataset, you could provide this prompt:
Create a page that analyzes our product sales by region and date.
Copilot will go to work. It identifies tables like 'Sales,' 'Products,' 'Geography,' and 'Calendar,' and populates a new page with a logical layout of visuals. You might get a map showing sales by state, a bar chart of top-selling products, a line chart of sales over time, and a few slicers for filtering. This baseline report is generated in seconds.
Refining Your Report with Follow-Up Commands
The real power lies in the follow-up conversation. You don't have to start over if the first result isn't perfect. You can treat it like you're giving feedback to a junior analyst running beside you.
Review the page Copilot created and ask for adjustments. For instance:
- "Add slicers for Product Category and Sales Rep."
- "Change the bar chart into a treemap to show product sub-category sales."
- "Use a darker theme for this page."
- "Add a card that displays total gross profit."
With each command, Copilot will update the report canvas. This interactive method lets you brainstorm and iterate far faster than you could by manually building and formatting each visual yourself.
Mastering DAX with Copilot
Writing DAX can be one of the steepest learning curves in Power BI. Copilot acts as an expert that can generate, explain, and debug measures for you.
Generating DAX Measures from Scratch
To use Copilot for DAX, go to the Quick measures button in the Home ribbon. You’ll find an option to "Create with Copilot". You can describe the calculation you need and Copilot will handle the syntax.
Let's say you need to calculate the average sale amount. Your prompt could be as simple as:
Average sale price from the Sales table
Copilot would generate the corresponding DAX:
AVERAGE('Sales'[SaleAmount])
Requesting Complex Calculations
This feature is a lifesaver for more complex, time-intelligence-based calculations that often cause headaches. Think about things like year-over-year growth, moving averages, or running totals.
Prompt example for a complex measure:
Calculate the year-over-year difference for total sales.
Copilot will recognize the need for functions like CALCULATE and SAMEPERIODLASTYEAR and produce the correct formula, complete with error handling.
Explaining Existing DAX Measures
Another incredible use case is learning. If you inherit a complex report or just want to understand someone else's work, Copilot can help. Simply create a measure, paste a complex DAX formula into it, then summon the chat pane and ask it to explain the formula for you. This deconstructs nested functions and confusing logic, making it a powerful tool for leveling up your own DAX skills.
Creating Quick Data Summaries with the Narrative Visual
Great reports don't just show data, they explain it. The "Narrative with Copilot" visual automatically generates text-based summaries for your data.
Free PDF · the crash course
AI Agents for Marketing Crash Course
Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.
How to Add a Narrative Summary
From the Report view, find and click the Narrative with Copilot icon in the Visualizations pane. This will add an empty text box to your canvas. A button inside prompts you to "Generate" a summary based on the data presented on the current report page.
Customizing the Narrative with Prompts
After generating a default summary, you can customize it from the prompt box within the visual itself. This gives you control over the tone, format, and focus of the text. Because the visual is interactive based on filters, the narrative will update as you slice and dice your data.
Here are a few powerful prompt examples:
- "Give me a bulleted list of the top 3 insights from this data."
- "Write this summary in a more professional tone for an executive audience."
- "Specifically call out which customer segment had the biggest drop in sales last quarter."
This is a fantastic way to quickly add an executive summary box or helpful metric descriptions to your dashboards.
Best Practices for Working with Copilot in Power BI
To get the most out of this new AI tool, keep a few best practices in mind:
- Use Clear Naming Conventions: Copilot understands your data model based on the names of tables and columns. A column named
CustomerSegmentis much easier for the AI to interpret thanCust_Seg_Grp_1. Clean, human-readable data models yield the best results. - Start Broad, Then Get Specific: For report creation, it's often best to start with a high-level goal and then provide specific follow-up commands to refine the output. This is usually more effective than giving one massive, highly detailed prompt.
- Always Verify the Output: Treat Copilot as a very fast and knowledgeable assistant, not an all-knowing authority. It's an excellent "first draft" generator, but you should always double-check the DAX measures and chart configurations it produces. Sometimes, it might misunderstand a relationship or write a formula that doesn't quite fit your business logic.
- Provide Context: When asking for measures or visuals, refer to the actual table and column names in your prompt (e.g., 'Sum of ProfitAmount from the Sales_Fact table'). The more context you provide, the better the output.
Final Thoughts
Copilot for Power BI is a massive step forward, transforming hours of manual work into seconds of conversation. Instead of getting bogged down in the technical minutiae, this AI partner lets you jump straight to analyzing results and discovering insights. Learning how to collaborate with it will fundamentally speed up your process from raw data to actionable reports.
The goal of AI in analytics is to make data easier for everyone to use and understand. At Graphed, we focus on this by allowing marketing and sales teams to create entire dashboards just by describing what they need in plain English. For teams who don't live inside tools like Power BI but still need instant, real-time answers from data sources like Google Analytics, Shopify, and Salesforce, using natural language to build what you need makes analysis accessible to everyone - not just the data experts. You can see how we're making that possible with Graphed.
Related Articles
How to Create Facebook Ads API for Claude Code
Create a Meta Marketing API access token that never expires, using a system user, so Claude Code can pull Facebook ad reports and manage campaigns.
ClickHouse vs Postgres: When to Switch to Columnar
Comparing ClickHouse vs PostgreSQL means matching storage architecture to workload. Row storage wins for transactions. Columnar wins for scan-heavy analytics.
Google Ads for Orthodontists: Get More Booked Consults
Most orthodontic Google Ads campaigns waste money on low-intent searches. The problem is structure, not the platform.