Episode 311 – Microsoft Ignite 2025: What Actually Matters for Fabric Pros
Microsoft Ignite has always walked a line between forward-looking vision and carefully crafted messaging. It’s rarely the place for deep technical dives—it’s where Microsoft lays out its next 18 to 24 months and unveils the features they want the world to notice. That’s exactly why we brought together three different viewpoints for this episode: Jason, John, and our guest Stephanie Bruno, Data Platform MVP and Data Witch. With so much landing at once, having multiple perspectives is the only way to cut through the noise.
Everyone came prepared with their own lists of standout Fabric announcements—Jason went for efficiency and prompted an AI tool, John assembled his top picks, and Stephanie curated her own set. As expected, the overlap was minimal. And that’s where the interesting conversations start.
Fabric Databases Go GA—A Bigger Shift Than It First Appears
One of the major headline announcements was Fabric Databases reaching general availability in both SQL and Cosmos DB flavors. The GA has been long anticipated, and with the feature now officially out, the implications are far clearer.
The session delivered by Anna was impressively full of content—she moved quickly, packing in a huge amount of detail about what Fabric Databases mean across operational, analytical, and AI workloads.
The real story is that this milestone gives organizations a fabric-native construct that spans transactions, analytics, and intelligence in the same place. It’s a conceptual unification Microsoft has been aiming toward for years.
Questions remain about how this shift affects traditional SQL workloads, especially for teams heavily invested in Azure SQL or on-prem SQL. Some features may land first—or only—inside Fabric, which encourages deeper adoption of the platform. That’s something enterprises will want to watch.
John highlighted a very different angle based on real-world project work he’s doing now: for scenarios requiring real-time operational and analytical capabilities in one place, Fabric SQL is already proving to be a perfect fit.
User-Defined Functions: The Most Practical Spark from Ignite
Stephanie pointed to User-Defined Functions (UDFs) as one of the most exciting announcements of Ignite—her “gateway” into exploring more of Fabric’s deeper capabilities. Several improvements landed, including:
- UDFs callable from Activator
- Expanded variable library support
- Integration with Azure Key Vault
These enhancements make UDFs useful far beyond their original scope.
The discussion naturally turned to how UDFs behave across different data sources. At this stage, they remain a Fabric-native capability; working with mirrored or shortcut data can introduce some limitations.
What opens the door wider is how UDFs connect to write-back scenarios. Jason noted that this aligns well with the growing opportunities for deeper interaction between Power BI, Fabric SQL, and Power Apps. The building blocks are there—even if the overall workflow spans multiple tools today.
He also mentioned a personal aside: his son Sam is now learning Python and may be recruited to help build Fabric demos this summer. With AI tools making Python even more accessible, the timing couldn’t be better.
The Fabric IQ Puzzle—And Why It Matters
Fabric IQ made its debut as part of the keynote’s big narrative. It’s an attempt to unify the various intelligence features across Microsoft’s data platform under one umbrella. It raised more questions than it felt like it answered—particularly around the role of ontologies and how they interact with semantic models.
John offered a clearer interpretation: Fabric IQ represents an intent to expose verified semantic models as authoritative knowledge sources for AI. In theory, that means better LLM interactions grounded in real data relationships.
Beyond branding questions, there’s an ongoing conversation about how capabilities shift between Power BI and Fabric’s real-time experiences. New features like graph visualizations appear in those dashboards first, which could signal a broader platform evolution. The hope is that innovation comes without sacrificing the strengths that users rely on today.
Fabric Data Agents in Teams: The Quiet Announcement with Huge Potential
While scanning through updates, Stephanie spotted a major addition tucked near the bottom of a blog post: Fabric Data Agents are becoming available directly inside Microsoft 365 Copilot in Teams.
This removes a major layer of friction. Previously, working with data agents required navigating Copilot Studio and juggling connection challenges. Now the story becomes far more natural—ask questions of your data where your conversations already happen.
The integration still has rough edges, but it’s a significant step toward natural, data-aware collaboration. John emphasized how meaningful this is because data agents understand context in ways LLMs alone cannot—they respect models, relationships, and structure rather than treating everything like an unstructured blob.
Jason connected this to long-standing conversations with friends in the AI world about how different natural language systems behave. His ideal future: ask a question like “Look at the last five sales periods by region and tell me where bicycles sold best,” have the system return an accurate insight with a visualization, and pin it directly to a report. This update moves that vision closer to reality.
More Highlights Worth Knowing
SharePoint and OneDrive Shortcuts
John pointed out one of the most practical quality-of-life improvements: you can now use shortcuts to access data in SharePoint and OneDrive directly, without needing to move files into Fabric. Structured files promote to tables automatically, while more complex ones may still need manual work—but the friction is dramatically lower.
Fine-Grained Write Access to One Lake
One Lake finally supports precise write permissions. Instead of granting broad workspace-level access, organizations can now manage permissions directly at the data layer—far closer to real-world needs.
Capacity Events Flow to Real-Time Hub
Real-time operational visibility got a boost. Capacity events now appear in the Real-Time Hub, making it possible to subscribe to throttling alerts proactively instead of discovering issues after the fact.
Event Streams Add HTTP Connector
With the new HTTP connector, Event Streams can call external APIs—like weather services—without writing code. John noted that this replaces patterns many organizations built with Azure Functions, dramatically simplifying pipelines.
Variable Libraries Mature
Jason highlighted how Variable Libraries have grown into essential infrastructure for CI/CD and DevOps workflows within Fabric. Their GA release with broader support reflects the platform’s maturation.
.pbir Becomes the Default Report Format
Though not specifically an Ignite announcement, the shift to the .pbir format modernizes Power BI file management. Jason credited Rui for consistently delivering strong updates in this area.
Where’s Power BI in All This?
A recurring observation: explicit Power BI content is increasingly scarce at Ignite. Searches turn up almost nothing, with focus shifting to Power Platform and Fabric instead. This trend isn’t new, and it mirrors patterns seen around SharePoint in past years.
Jason’s read is that Microsoft is emphasizing Fabric’s broader story over Power BI’s individual brand. But Power BI remains essential in real-world usage. His approach today? Use Fabric as the powerhouse backend, Power BI as the polished frontend, and Copilot as the bridge for summarization and Q&A. It’s still a highly effective combination.
Fabric SQL Ecosystem and Governance Advances
Stephanie highlighted the new Spark connector for Fabric SQL Database—a major quality-of-life improvement for notebook scenarios. Direct reads and writes remove a long-standing limitation and streamline advanced workflows.
She also pointed to the expanding governance experience. The One Lake Catalog admin report now spans multiple workspaces, and deeper navigation is available with visuals like decomposition trees. For organizations running large Fabric environments, this increased visibility is essential.
December: The Perfect Month to Experiment
Ignite landing in November creates a rare advantage—December is naturally quieter, giving teams time to test new capabilities before the new year. Even with a busier-than-normal season, there’s still room for experimentation.
Stephanie added that Fabric Data Days are offering free DP-600 and DP-700 exam vouchers through December 31st, along with contests and community activities. It’s an easy opportunity for anyone looking to grow their skills.
The Bottom Line
Ignite’s announcements always blend innovation, vision, and practical updates. This year’s mix—Fabric Databases, Data Agents in Teams, fine-grained security, Event Streams, and more—provides clear value for organizations adopting Fabric strategically. Some areas, like Fabric IQ, are still taking shape, and that’s part of the natural evolution of a platform this ambitious.
The real clarity comes from conversations—like the one between Jason, John, and Stephanie—where different perspectives spark new understanding. That synthesis is where the value lies, more than any single announcement.
Links & Resources
- Episode 310 – October 2025 Feature Summaries
- Fabric Databases— a unified, SaaS-native experience for modern data workloads (Generally Available)
- Announcing SQL database in Fabric (Generally Available)
- What’s New for Fabric Data Agents at Ignite 2025: Unlocking Deeper Data Reasoning and Seamless AI Interoperability
- BIFocal Podcast Website
Subscribe: SoundCloud | iTunes | Spotify | TuneIn | Amazon Music
