Graphed LogoGraphed
Login

How to Use Quick Analysis in Power BI with AI

Cody Schneider•

Buried deep within your company's data are the answers to your most pressing questions - Which campaign drove the most sales? Where are our leads coming from? Why did last month's revenue dip? The challenge has never been a lack of data, it's been turning that data into clear, quick answers without spending hours building complicated reports. Power BI's Quick Analysis feature is designed to bridge that gap using AI and natural language.

This article shows you how to use this feature to chat directly with your data. We'll walk through exactly how to get started, use it to pull quick insights, and create shareable visuals in seconds instead of hours.

What is Quick Analysis in Power BI?

Quick Analysis is Microsoft’s response to a common problem: Traditional business intelligence dashboards are powerful, but they require a significant learning curve. You need to understand data models, relationships, and how to drag and drop different fields, measures, and filters to get the answer you’re looking for. For non-technical team members, this can feel like trying to pilot a rocket ship just to drive down the street.

Instead of manually constructing visuals, the Quick Analysis feature allows you to type your questions in plain English. You can ask things like, "What were our total sales last quarter?" or "Show me revenue by product category as a column chart." Power BI's AI interprets your request, analyzes the underlying dataset, and instantly generates a visual to answer your question. Think of it as having a junior data analyst available 24/7 who can build visuals for you on command.

This approach dramatically lowers the barrier to entry, making data accessible to everyone on your team, not just those with "analyst" in their job title.

Getting Started with Quick Analysis in Power BI

Using the Quick Analysis feature, officially called "Suggest a visual with Copilot," is straightforward, but your data needs to be properly set up in Power BI Desktop first. Here’s a simple step-by-step guide to get you going.

Note: This feature is part of Microsoft Fabric and typically requires an F64 or P1 capacity for use. Your administrator will need to enable Copilot settings in Fabric for your organization.

1. Prepare Your Data Model

For the AI to give you accurate answers, it needs to understand your data. Before you start asking questions, make sure your report's data model in Power BI is clean and logical. This means:

  • Your tables have clear, understandable names (e.g., "SalesData" instead of "tbl_sl_2023_final").

  • Relationships between tables are correctly defined. For example, your "Sales" table should be properly linked to your "Products" table.

  • Column names are intuitive (e.g., "CustomerName," "SaleAmount," "OrderDate").

Good data preparation is the foundation of any reliable analysis, AI-powered or not. A well-organized model allows the AI to correctly interpret what you're asking for.

2. Open a Report in Power BI Desktop or Service

You can use the AI feature in both Power BI Desktop and the Power BI Service. Open an existing report that you want to analyze. Once it's loaded, you’ll find the button to launch the analysis pane.

3. Locate the 'Suggest a visual' Icon

Look for the visual icon in the ribbon at the top of the interface. This button launches the Copilot pane, where you can start interacting with the AI. Clicking it opens a conversational interface right alongside your report.

From here, you can start typing your questions directly into the prompt box to generate insights.

4. Describe the visual you want to create

This is where the magic happens. Start asking your questions just like you would to a team member. You don’t need to know technical terms or query languages. You can be as simple or as specific as you want.

From here, you can give a prompt to create your visual. The most basic prompt format is: "(what data you want on your x-axis) by (what data you want on your y-axis)." But don't worry about always being perfectly formatted, since Copilot can extract and enrich what you're trying to accomplish based on the data you're analyzing. Let’s walk through a few examples of how this would be done in practice by analyzing customer lifetime value against customer acquisition cost (CLV:CAC Ratio).

As you use the tool more, you will start to understand its nuances. Once you finalize the new chart, you can change the aggregation type or even add it to a brand new dashboard page:

Drilling Down Further with Follow-up Questions

The real power of conversational AI tools isn’t just asking one-off questions, it’s their ability to explore data through follow-up prompts. Once Power BI generates an initial visual, you don’t have to start from scratch to tweak it. You can refine your analysis iteratively.

For example, after creating the "Sales by Region" pie chart, you might ask follow-up questions like:

  • "Now show that as a bar chart."

  • "Filter this to only include the last quarter."

  • "Add the profit margin to this chart."

  • "Sort by highest performing region first."

Each command builds on the last, letting you "peel back the layers" of your data and uncover deeper insights without having to navigate through menus and filter panes. This process of iterative querying enables a true conversation with your data, letting your curiosity guide the analysis. If an insight raises a new question, you can immediately pivot and explore it.

Tips for Writing Effective Prompts

While the AI is designed to understand casual language, you can get better, faster results by following a few simple best practices. You don't need to be an expert prompter - a simple, incomplete thought can often work - but providing clarity will always improve accuracy.

  • Be Specific With Metrics and Dimensions: Instead of "Show me our performance," which is ambiguous, ask for "Show me our revenue by product category for the last 90 days." Explicitly mentioning metrics (revenue), dimensions (product category), and timeframes (last 90 days) leaves little room for misinterpretation.

  • Ask for a Specific Chart Type: If you have a specific visualization in mind, ask for it. Phrases like "...as a line chart" or "create a scatter plot showing..." guide the AI to generate the exact visual format you need for storytelling.

  • Use Clear Language: Use the same names that appear in your data model’s columns and tables. If your sales column is named "TotalRevenue," using that term in your prompt can improve accuracy. The AI is smart, but it's not a mind reader!

  • Start Simple and Iterate: If you have a complex question, start with a broader request and use follow-up prompts to drill deeper. For example, start with "Total sales this year" and then narrow it down with "by month," then "for the top 5 products."

Getting proficient with this tool is less about learning a new software and more about honing your ability to ask clear questions. It empowers anyone, particularly junior team members or those in non-technical roles, to become more data-driven.

Limitations and When to Use Manual Analysis

AI-powered analysis is incredibly powerful for speed and accessibility, but it's not a complete replacement for manual dashboard creation. It works best for getting quick answers, exploring new datasets, or building one-off visuals for a presentation. When you see something interesting in your AI-generated chart, you use your domain experience to draw conclusions about why that insight might be significant.

For standardized, highly-formatted company dashboards that require pixel-perfect layouts, specific branding, complex DAX measures, or row-level security, the traditional report builder in Power BI is still the better tool. Think of Quick Analysis as a fast, agile speedboat for data exploration, while the full Power BI report editor is the battleship you use to build robust, permanent reporting assets.

Final Thoughts

All in all, Quick Analysis with Power BI is one small step on our journey towards a more intelligent data solution. Getting a simple chart or number out of a data warehouse should not take long with new and improved large language models (LLMs). This functionality will enable new groups that are unfamiliar with data and how to present solutions to their companies effectively communicate the needs of their departments or customers without becoming a programmer in the process.

For organizations already committed to the Microsoft ecosystem and who primarily rely on data stored within a dedicated warehouse, Power BI’s Quick Analysis is an emerging tool. However, if your data is scattered across platforms like Google Analytics, Shopify, Facebook Ads, and Salesforce, getting everything in one place can be a struggle. We built Graphed for precisely this reason. We offer one-click integrations to connect all your sources in seconds, then let you use natural language to build real-time dashboards and reports, so you can stop manually pulling CSVs and spend your time acting on insights instead.