Chat With Your Data - What Is It?

In this video, Pragmatic Works trainer Greg Trzeciak explores one of the fastest‑evolving AI tools within the Microsoft ecosystem: Copilot’s Chat With Your Data. As organizations adopt AI-driven workflows, many data professionals are asking an important question — is Copilot replacing the data analyst? Greg walks viewers through the current state of Copilot, its rapid evolution, and what its capabilities mean for the future of analytics.

How Copilot Has Evolved

Greg begins by reflecting on how Copilot originated as a general-purpose AI assistant similar to ChatGPT. At first, users asked fun or simple prompts such as recipe suggestions or workout plans. Over time, Copilot became deeply integrated across Microsoft 365 tools like Outlook, Excel, Teams, PowerPoint, and Office applications, assisting with emails, summaries, formulas, design ideas, and more.

As organizations adopted Copilot licenses, employees learned to use more advanced prompting — generating summaries, creating images, writing documents, and layering in Graph data. Greg explains that as the AI expanded, its speed and capabilities pushed organizations to embrace Copilot more seriously as part of daily work.

Introducing “Chat With Your Data”

The major shift came with Copilot’s integration into Power BI. Greg explains that Copilot was originally a helpful companion inside Power BI Desktop and Power BI Service, allowing users to ask questions of a report and receive insights. But now, Chat With Your Data operates at full scale:

• Users can query any semantic model they have access to.
• Copilot can generate visuals, run DAX queries, and retrieve existing report visuals.
• The experience functions like a “Google for your data,” directly inside the Power BI Service.

Greg highlights that this isn’t just an enhancement — it’s a transformative experience that reshapes how organizations consume insights.

How Chat With Your Data Works

To demonstrate the process, Greg loads a semantic model used in a Dashboard in a Day class: a dataset tracking U.S. bank failures. He selects the dataset through the Copilot pane and asks a natural-language question:

“What are the top five states by bank failures?”

Copilot responds immediately with:

• A list of the top states
• A linked reference to the original report
• A Chicago‑style annotation showing the owner, workspace, and refresh time
• The full original visual from the Power BI report

Greg notes something important: Copilot could have generated a new table or DAX result, but instead it recognized that a relevant visual already existed and reused it. This demonstrates that Copilot is designed to strengthen—not replace—the work data analysts already do.

What About Custom Visuals?

Greg confirms that custom visuals are still supported. Every visual a data analyst creates — whether default or custom — remains valuable. Copilot can:

• Surface these visuals during a query
• Generate alternative visual types
• Allow users to explore and reshape results

Users can expand visuals, convert them to other formats (such as tables), or dive deeper using the “Explore” feature. Greg stresses that although this feature is often overlooked, it becomes significantly more powerful when combined with Copilot.

Will Copilot Replace Data Analysts?

Greg addresses the most common fear directly: No — Copilot is not replacing the data analyst. Instead, it raises the bar for the skill sets analysts should develop.

He encourages analysts to focus on areas such as:

• Preparing data for AI (clean models, relationships, business logic)
• Adding synonyms so Copilot understands organization‑specific terminology
• Defining verified answers so Copilot knows the correct source of truth
• Improving semantic models to support better natural‑language queries

Copilot is only as effective as the data model behind it. A poorly modeled dataset will lead to unclear answers, confusion, or inaccurate results. Analysts play a critical role in shaping the model so that Copilot becomes a well‑informed AI collaborator rather than a “rookie analyst” making guesses.

The Future: Preparing Your Organization

Greg emphasizes that organizations across the country — and internationally — are moving into Fabric and adopting Copilot-driven workflows. The demand for AI-ready data has never been higher. Pragmatic Works now offers services to help organizations:

• Prepare data models for AI and Copilot
• Enable Chat With Your Data for enterprise use
• Modernize data estates for Microsoft Fabric
• Train teams on effective prompting and Copilot adoption

Greg closes by reminding viewers that AI isn’t going anywhere. Chat With Your Data is here today, and data teams should embrace it to stay ahead.

Stay pragmatic — and get ready for 2026.

Don’t forget to check out the Pragmatic Works On-Demand Learning platform for more training on Power BI, Copilot, Fabric, and AI readiness.

Don't forget to check out the Pragmatic Works' on-demand learning platform for more insightful content and training sessions on [TECHNOLOGY DISCUSSED IN BLOG POST] and other Microsoft applications. Be sure to subscribe to the Pragmatic Works YouTube channel to stay up-to-date on the latest tips and tricks.