Power BI – March 2025
Version Updates
During Fabcon 205 the latest release for Power BI was released. Compared to the last 2 months there are a lot of new features to explore.
General
1. Power BI Desktop start-up performance improvements.
Starting from the March release there will be a significant improvement of the start up of Power BI Desktop. For example, opening an empty report will go from ~8 seconds to ~4. And opening a PBIX report would go from ~17 to ~14 seconds. So in general you’ll save a couple of seconds when starting up a report.
AI
1. Copilot Summary in Teams/Outlook.
A new button has been added that gives you a Copilot summary for the Power BI report you’re looking at when pressed.
2. DAX Queries by Copilot using user-created hierarchies.
Copilot can now use user-created hierarchies as a reference when generating a DAX query.
3. Language improvements for Copilot.
A new LLM has been added to Copilot to help it understand common synonyms, phrasings, and individual values. For example, in the past it had issues understanding what ‘Export’ means in the context of suppliers and products. As a result more questions can be answered by Copilot.
4. Ad-Hoc calculations
Previously, when a question required Copilot to perform calculations based on the model, Copilot was not able to provide an answer. For example, if we had a data model with employee data and a hired date. Copilot was not able to answer “How many employees were hired before 2020”. But now that it is able to perform some calculations itself (through DAX) it is able to answer this questions. And, you can always verify what DAX was used to get to an answer.
Reporting
1. Annotations in PowerPoint.
You are now able to add annotations to Power Bi visuals that are displayed in PowerPoint. Annotations will stay visible as long as the data point it is attached to exists. If the data point does not exist anymore or when it’s filtered out, the annotation will not be displayed.
2. Visual Calculations Improvements.
The trend of the previous updated continues. The visual calculations gets more improvements with this update. Similar to Excel, columns will be highlighted to show that they are being used in the calculation.
3. Reference Line improvements.
- Shade areas: you can now add a shade above or below your reference line.
- Y-Axis reference lines for Line & stacked column and Line & Clustered column charts
4. Card visual improvements.
While the card visual is still in PREVIEW, more improvements are being added.
- Categories (small multiples) get 2 new styles:
- Table (Data points appear in a tabular format, similar to a traditional spreadsheet)
- Card (Data points are displayed as individual cards)
- Headers can now be further customized by changing the location and orientation of the headers.
- Conditional formatting for each category (small multiple)
5. Style presets!
A new formatting option has been added to visuals. You can now predefine multiple formatting options in a custom theme file. And these can be applied to a visual. This will make it easier to apply multiple formatting changes to a visual, eliminating the need to manually change each option.
Modelling
1. Get Object Names (PBIR) from your report.
In PBIR each element (visual/page/bookmark/…) is saved in it’s own folder. The name of this folder is a generated ID so it’s not always easy to see what folder belongs to which element. When enabling this feature, it will be possible to right click any element in a Power BI report and get the object name (ID for the folder in PBIR). This will make it a lot easier to develop/understand your PBIR.
2. Direct Lake improvements:
- Create Semantic models in Power BI Desktop using Direct Lake. (PREVIEW)
- TMDL view can be used to edit your semantic model in Direct Lake Mode (PREVIEW)
- Edit Direct Lake Semantic models in Power BI Desktop from web (PREVIEW)
3. Notebooks for your semantic model. (PREVIEW)
You will now be able to use (Fabric) notebooks with your semantic model. There are 3 types of notebooks you can use:
- Best Practices Analyzer (offers tips to improve the design and performance of your semantic model)
- Memory Analyzer (show you memory/storage statistics about the objects in your semantic model)
- Community Notebooks (notebooks created by the community)
