Microsoft just flipped the script on how we use data. If you’ve been ignoring the incremental updates over the last few months, today is the day to stop. Power BI isn’t just a "chart maker" anymore. It’s becoming an agentic AI hub, and honestly, some of the legacy features you probably rely on are officially on the chopping block.
The biggest news hitting the wire right now? The death of Q&A.
Microsoft confirmed that the classic Q&A visual, that old-school natural language box we’ve used for years, is being deprecated. It’s moving to the "legacy" pile to make room for a much more aggressive rollout of Copilot. If your reports depend on those specific Q&A buttons, you have a deadline. By December 2026, they’re gone. It sounds like a long way off, but for enterprise-level reporting, the migration needs to start basically now.
Copilot is Writing Your DAX Now (For Real)
Let's talk about the DAX Query View. If you’ve ever stared at a screen for three hours trying to figure out why a CALCULATE function isn't respecting a filter context, you're going to love this. Copilot for DAX is now generally available.
It’s not just a chatbot sitting on the side. It actually lives inside the DAX Query View. You can type, "Give me a measure for year-over-year growth but exclude the promotion periods in Q3," and it spits out the code.
Is it perfect? No.
Is it better than scouring Stack Overflow? Absolutely.
✨ Don't miss: New DeWalt 20V Tools: What Most People Get Wrong
The cool part is how it handles the "why." It doesn't just give you the code; it explains the logic. This is huge for teams where maybe one person is a DAX wizard and everyone else is just trying to keep their head above water.
What about the "ARM" thing?
This is a niche bit of Power BI news today that actually matters for performance. Power BI Desktop now runs natively on Windows on ARM. If you’re using one of those new Surface Pro models or a high-end Snapdragon laptop, you’ve likely noticed the lag when Power BI emulates x64. That’s over. It’s faster, the battery doesn’t melt, and it feels like a real app instead of a resource hog.
The Map Migration: Bing is Out, Azure is In
If you open your reports today and the maps look… different, or if you’re getting warnings, here is why. Microsoft is killing off the Bing Maps visual.
Everything is moving to Azure Maps.
It’s a forced migration, but honestly, it’s a win. Azure Maps is way more scalable and handles high-density data points without the weird stuttering Bing used to have. If you have legacy reports with the old globe icon, you’ll need to manually switch the visual type to Azure Maps. Most of the time it’s a one-click fix, but keep an eye on your custom formatting; sometimes the heat maps need a little nudge to look right again.
🔗 Read more: Memphis Doppler Weather Radar: Why Your App is Lying to You During Severe Storms
Fabric is Eating Power BI (In a Good Way)
You can't talk about Power BI news today without talking about Microsoft Fabric. The two are becoming inseparable.
The latest update introduced Dataflows Gen2 export. This is a game-changer for people who hate "shadow BI." You can now take the data you cleaned in Power Query and export it directly into a Fabric Lakehouse or a Warehouse.
Think about that for a second.
Previously, your "clean" data lived and died inside a single Power BI dataset. Now, you can treat your Power BI cleaning steps as a legitimate ETL process for the whole company. It turns Power BI into a data producer, not just a consumer.
Direct Lake is the New Import Mode
We’re seeing a massive shift toward Direct Lake mode. If you’re dealing with massive datasets—think millions or billions of rows—Import Mode starts to crawl. Direct Lake lets Power BI read directly from OneLake (the Fabric backbone) with the speed of Import Mode but without the "refresh" headache.
💡 You might also like: LG UltraGear OLED 27GX700A: The 480Hz Speed King That Actually Makes Sense
It’s the closest thing to "real-time" we’ve ever had without the performance penalty of DirectQuery.
Small Tweaks That Actually Matter
- Button Slicer is GA: The new button slicer is finally out of preview. It supports "hero images," meaning you can make your slicers look like a professional website menu instead of a list of checkboxes.
- Mobile Copilot: You can now use Copilot on the mobile app. You can literally ask your phone, "Why did sales dip in Georgia this morning?" and it will scan the semantic model and give you a summary.
- Git Integration: For the developers out there, the Git integration is getting more stable. You can finally track version history for your
.pbipfiles without it being a total nightmare.
Real-World Bottlenecks to Watch Out For
Look, it’s not all sunshine. The biggest frustration right now is Copilot licensing. To get these new AI features, your organization needs to be on a Fabric F64 capacity or higher, or have Power BI Premium (P-SKUs).
If you’re on a Pro license, most of this "cool stuff" is locked behind a paywall.
Also, the deprecation of Metric Sets is causing some friction. Microsoft wants everyone using "Scorecards" and "Metrics" in the Power BI service, but the transition hasn't been perfectly smooth for companies that built their whole KPI strategy around the old Metric Sets.
Your "To-Do" List for This Week
Don't just read the news; act on it. If you’re managing a Power BI environment, here are the three things you should do before Friday:
- Audit your Maps: Open your most-used reports. If you see a Bing Map visual, swap it to Azure Maps now. Don't wait for it to break.
- Test the DAX Query View: Stop writing measures in that tiny little formula bar at the top. Open the DAX Query View (the icon on the left) and try the Copilot "Explain" feature. It’ll change how you learn the language.
- Check your Q&A usage: Go to your Power BI Service usage metrics. Are people actually using the Q&A box? If they are, start training them on how to use the "Chat with data" feature in the Copilot pane instead.
Power BI is moving fast. If you stay on the 2023 version of your workflow, you’re basically working twice as hard for half the result. The integration with Fabric and the move toward agentic AI means we’re moving away from "building reports" and toward "curating data for AI to explain." It’s a weird shift, but it’s where the industry is heading.