Episode 294 – Power BI & Fabric April 2025: Feature Updates That Actually Matter

Jason and John sat down (early, much to Jason’s chagrin after a rare sleep-in opportunity) to power through both the Power BI and Microsoft Fabric April 2025 feature summaries in a single episode. April brought a lighter update cycle than March’s feature deluge, but the duo found plenty worth discussing—from game-changing desktop capabilities to quality-of-life improvements across the platform.

Copilot Goes Mainstream

The biggest news for everyday users? Copilot features are no longer locked behind F64 capacity requirements. As long as you’re running an F2 or higher (which most organizations can afford), you’ve got access to AI-powered assistance in Power BI. This democratization means those copilot sections in the feature summaries finally matter to the masses.

John shared a practical tip: when working with trial capacities that lack copilot support, you can point Power BI Desktop at a separate F2 capacity just for copilot work. This keeps your AI queries from bogging down compute-intensive operations running elsewhere—a smart architectural move for managing capacity resources.

Desktop Gets Direct Lake Editing

The modeling section delivered April’s headline feature: live editing of semantic models in Direct Lake mode using Power BI Desktop. Previously, Direct Lake models could only be created and modified through the service. Now desktop’s superior tooling—including Tabular Editor View—works with these models.

The catch? You’re working live against the service, not publishing changes. It’s similar to live-connecting to a semantic model for reporting, except now you can actually edit the model itself. John noted the UI still needs refinement (you have to start by creating a “new report” even though you’re not creating one), but the capability represents a significant workflow improvement.

The feature requires XMLA read/write permissions at the workspace level, meaning you’ll need Fabric or Premium capacity. Access it through the OneLake Catalog by selecting a semantic model and choosing “Edit model” from the dropdown—easy to miss if you don’t know where to look.

Mobile Views Get Smarter

Power BI’s mobile view editor now includes an auto-layout button that generates optimized mobile layouts automatically. Jason, a self-described mobile view evangelist who creates them for 75% of his reports, tested the feature and found it produces a solid 60% starting point—good enough to save significant time even if you need to tweak the results.

For those who couldn’t be bothered creating mobile views before (John estimates only 5% of reports have them), this one-click option removes the friction. Your mobile users will thank you.

The New File Picker Has Limits

The updated file picker in Power BI Desktop is now enabled by default, offering excellent OneDrive and SharePoint integration. But John discovered an important caveat for CICD workflows: you can’t use the picker to save PBIP or PBIT files directly to OneDrive or SharePoint. You have to save locally (or to a synchronized OneDrive folder) instead.

Jason pointed out this actually makes sense—when you’re using Git for version control, you want files saved locally where your Git client can track them, not syncing through multiple services simultaneously.

Data Wrangler Comes Out of Hiding

Perhaps the most impactful change for data scientists and engineers: Data Wrangler is now prominently featured in the notebook ribbon, just three buttons from Copilot. Jason was enthusiastic about this visibility boost—Data Wrangler has existed for over two years as “Power Query for notebooks,” but its previous obscurity meant many users never discovered it.

The tool itself gained new capabilities too: rule-based AI suggestions, natural language instructions for transformations, and automatic conversion between Pandas and PySpark code. Combined with improved discoverability, Data Wrangler should finally get the adoption it deserves.

Real-Time Analytics Keeps Evolving

The RTI section brought several notable updates:

• Azure Monitor Integration: Log Analytics and Application Insights now work as first-class citizens in KQL Query Sets, eliminating the convoluted workarounds previously required
• Event House Enhancements: The main Event House screen continues its rapid evolution with better ingestion monitoring, filterable time periods, and improved data exploration tools
• Weather Connector: A new real-time weather data connector provides an easy entry point for learning RTI capabilities—no hardware required, unlike John’s beloved weather stations

John particularly emphasized the weather connector as a teaching tool. Unlike replaying old taxi or stock data, it provides genuinely real-time information that updates every minute, making it ideal for workshops and demonstrations.

What’s Coming

With Microsoft Build happening May 19-22 (free online registration available), John and Jason expect the May feature summary to pack more punch. They might even record that episode together in Germany during their upcoming workshop tour.

For now, April’s updates represent steady progress across the platform—not flashy, but meaningful improvements to daily workflows. As Jason noted about AI-generated features, even 60% completion saves substantial time. April delivered that and more.

---

Links:

• Power BI April 2025 Feature Summary
• Microsoft Fabric April 2025 Feature Summary
• Episode 293 – Microsoft Fabric March 2025 Feature Summary (Part 2)
• Microsoft Build

Subscribe: SoundCloud | iTunes | Spotify | TuneIn | Amazon Music