Power BI November 2025 Feature Summary

In the last year, every month additions have been made by Microsoft to the Copilot experience in Power BI. This month is no exception: standalone Co-pilot is now added to the mobile app, improvements have been made to “chat with your data” and some much requested improvements to verified answers have been made.  

Copilot and AI

Standalone Copilot in the Power BI mobile app (Preview)

You can now ask questions about your data from your mobile home screen. Copilot replies with instant insights and can even create visuals for you. You can interact with the visuals or share your findings with others. Voice dictation is supported on iOS for hands free insights on the go. 

Better Copilot for report creation

Copilot now understands intent with more detail and selects the right visuals more accurately. It also supports a wider range of visuals and gives more relevant suggestions based on recently opened content. 

Improved Verified Answers

Verified answers now respect the full visual state including slicers cross filters and drill through actions. More filter combinations are supported for complex questions and cards and Azure maps are now included. 

Remote Model Context Protocol Server for Power BI (Preview)

You can build custom AI agents that query your semantic models securely through Microsoft Entra ID. This opens the door to natural language driven analysis experiences directly on model data.

App Copilot (Preview)

Users can chat with curated app content to find reports and insights faster within that app experience. A copilot button will be added on the top of your report list in the app, allowing the user to chat with data from the underlying reports.  

Remote Model Context Protocol Server for Power BI (Preview)

Soon, the Remote MCP (model Context Protocol) Server will support chat with your data.  

This allows you to build custom AI agents that query your semantic models securely through Microsoft Entra ID. This opens the door to natural language driven analysis experiences directly on model data.  

Reporting

Automatically expand matrix columns

Tables and Matrixes are among the visuals with the most change requests among the core visuals of Power BI. It looks like Microsoft is slowly starting to address this, last month the Tables got a “Grow to fit” option, and now this is added for matrices too.  

The Grow to fit option distributes visual space evenly so a matrix fills the entire width without manual resizing. This will save you quite some time resizing, and is a step in the right direction of hopefully more additions to the core visuals in the near future.   

Card visual now generally available

The (new) Card visual has been in public preview for over 2 years, and is now generally available! The visual was already widely used, and has now also received hero images and collage layouts for creative storytelling. It gives you a more expressive way to present KPIs and branding elements in reports. 

Be aware though: There are some changes in settings between the preview version (of before November 2025) and general available version of this visual. You will probably have to change space settings to make your reports look good again: more info here on MS Learn.  

Enhanced image visual

More flexibility for images including: 

• Upload URL or DAX generated images

• Styling options like borders shadow title and shaped corners

• Hover or pressed interaction states 

• SVG support without workarounds 

And last but not least, you can now finally update an image of a visual instead of deleting it and having to create a new one.  

Modeling

Power BI Modeling MCP Server (Preview)

With the Power BI MPC Server, and it’s extension in Visual Studio Code, you can built a semantic model using AI agents. This makes use of the TMDL semantic model format, which allows you to do any change in the semantic model in code instead of the standard drag and drop interface of Power BI desktop. While that is already powerful in itself, the addition of an MCP and a VS code extension brings powerful new options, like automatically documenting the semantic model, creating translations, and refactoring measures into Calculation groups.  

This feature is in it’s early stages of public preview though and only time will tell how this will develop further.  

Semantic model version history now available

You can restore earlier versions of your model directly inside Power BI. Automatic version capture gives you confidence to experiment and correct mistakes easily. This is however, only available for semantic models edited in the web, or Direct Lake models that are live edited in Power BI desktop. Up to 5 versions per semantic model are saved.  

TMDL Visual Studio Code extension enhancements

For advanced semantic modellers, TMDL documents within PBIP Power BI projects are a way to change your semantic model using code.  

Semantic highlighting for DAX and Power Query, better navigation and formatting and localisation improvements for modeling as code workflows. 

General announcements

Deprecation of R and Python visuals in embedded solutions

Starting May 2026, Power BI services that embed content for external customers will no longer show R or Python visuals. They will load as blank instead. Now is the time to prepare alternatives if you rely on embedded analytics.

Power BI is moving toward a future where insights are faster to discover and easier to share. AI is becoming more helpful and visuals more polished. Modeling and governance are getting stronger to support enterprise growth.

If you want guidance on rolling out Copilot responsibly or improving embedded reporting, Bitmetric can support you every step of the way.

This overview is based on the official Microsoft Power BI November 2025 update by Adam Saxton.
👉 Power BI November 2025 Feature Summary

Let’s connect!

More from the Bitmetric team