Building a Dashboard with Copilot
You have clean data, a sound model, and a set of measures. Now comes the payoff: turning all of it into a one-page dashboard people can read at a glance. This is the visual layer, and it is where AI feels almost magical. Copilot can generate an entire report page from a sentence, and even without Copilot, AI can hand you a complete layout plan you assemble in minutes.
In this lesson you will build a real dashboard. You will learn how to prompt Copilot for a page, how to plan a layout with a free assistant, which visuals to use for which questions, and how to arrange everything so it tells a clear story.
What You'll Learn
- How to generate a report page with Copilot from a single prompt
- How to get a full layout plan from a free assistant if you lack Copilot
- Which chart type answers which kind of question
- How to arrange visuals so the dashboard reads top to bottom, left to right
Start With the Question, Not the Chart
Beginners often start by picking a pretty chart. Professionals start with the question the dashboard must answer. Before building, write one sentence: "This dashboard shows how revenue is trending and which regions and products drive it." Everything on the page should serve that sentence. Ask AI to help you focus:
"I'm building a Power BI sales dashboard for a manager who has 30 seconds to look at it. My measures are Total Revenue, Profit Margin %, Revenue YoY %, and Total Orders. What are the 4 to 6 most important things to show, and in what visual for each?"
You will get a prioritized plan, which becomes your blueprint.
Generating a Page With Copilot
If you have Copilot in Power BI, open the Copilot pane and describe the page:
"Create a report page titled 'Sales Overview'. At the top, show cards for Total Revenue, Profit Margin %, Revenue YoY %, and Total Orders. Below them, a line chart of Total Revenue by Month. Beside it, a bar chart of Total Revenue by Region. At the bottom, a table of the top 10 Products by Total Revenue. Add a slicer for Region."
Copilot assembles the page. It will not be perfect, expect to resize, recolor, and fix a title, but it delivers 90 percent of the work instantly. Review each visual: does it use the right measure? Is the aggregation correct? Adjust anything off, and ask Copilot to refine: "Change the Region chart to show Profit Margin % instead" or "Sort the product table descending by Revenue."
No Copilot? Get a Layout Plan Instead
Without Copilot, AI still saves you enormous time by planning the page. Ask:
"Design a one-page Power BI sales dashboard layout. I have these measures: Total Revenue, Profit Margin %, Revenue YoY %, Total Orders. I have dimensions Region, Product, Category, and a Date table. Give me a specific layout: what goes in the top row, middle, and bottom; which chart type for each; and which field goes on each axis. Explain each choice in one sentence."
You will get instructions like: "Top row: four KPI cards. Middle-left: line chart, Month on the X axis, Total Revenue on the Y. Middle-right: bar chart, Region on the axis, Total Revenue as value. Bottom: table of Products sorted by Revenue with a Profit Margin % column. Add slicers for Region and Date." Then you build each visual by hand, choosing the chart from the Visualizations pane and dragging fields onto it. It is slower than Copilot but completely free and teaches you the tool.
Choosing the Right Visual
Match the chart to the question:
- Card — a single big number. Use for KPIs like Total Revenue.
- Line chart — change over time. Use for trends by month or quarter.
- Bar or column chart — compare categories. Use for revenue by region or product.
- Table or matrix — exact numbers and multiple columns. Use for top-N lists and detail.
- Pie or donut — parts of a whole, but only with a few categories; bars are usually clearer.
- Map — values by geography, if you have country or region data.
- Slicer — an on-page filter (a dropdown or buttons) so viewers explore.
Unsure which to pick? Describe your question: "I want to show how each product category's share of revenue changed from last year to this year. Which Power BI visual is clearest, and why?" AI will recommend and explain, and often warn you away from a misleading choice like a 3D pie.
Arranging for Readability
People read a dashboard the way they read a page: top to bottom, left to right. Put the most important summary, your KPI cards, across the top. Place the main trend just below on the left, where the eye lands next. Supporting detail (tables, breakdowns) goes lower and to the right. Slicers go along the top or left so users see how to filter before they dig in.
Keep it calm: consistent colors, aligned edges, and no more than five or six visuals per page. A cluttered dashboard hides its own message. Ask AI for a critique: "Here's my dashboard layout: [describe the visuals and positions]. As a data-viz expert, tell me if the arrangement is clear for a busy manager and suggest up to three improvements."
Test It Like a Viewer
Click your slicers. Does the whole page update sensibly when you select one region? Do totals recalculate? This is a great final check that your measures and relationships all work together. If clicking a region changes some visuals but not others, a relationship may be missing, paste the symptom to AI to diagnose. When every visual responds correctly to a slicer, your dashboard is genuinely interactive, not just a static picture.
Key Takeaways
- Start from the question the dashboard must answer, then choose visuals to serve it.
- Copilot can generate a full report page from one detailed prompt; expect to refine, and ask it to adjust specific visuals.
- Without Copilot, AI gives you a precise layout plan (what, where, which chart, which field) that you build by hand for free.
- Match chart to question: cards for KPIs, lines for trends, bars for comparisons, tables for detail, slicers for filtering.
- Arrange top to bottom and left to right, keep it to five or six clean visuals, and test interactivity by clicking your slicers.