With another March nearly behind us, Microsoft used FabCon 2026 to showcase its latest innovations—and unsurprisingly, AI was front and center. But beyond the buzz, the message was clear: powerful AI starts with a strong data foundation. Atlantic had the privilege of attending the conference both as a vendor and an attendee, and we came away with valuable insights on the biggest announcements, emerging trends, and what they could mean for the future of data and analytics. In this post, we’re sharing the updates that stood out most and our thoughts on where Microsoft is headed next. 

The Big Theme from FabCon 2026 

If there was one message repeated across nearly every announcement at FabCon this year, it was this: AI is only as valuable as the data platform behind it. Microsoft’s strategy is clearly focused on making Fabric the unified foundation for analytics, governance, automation, and AI-driven experiences. 

That showed up in a few major ways: 

• Continued investment in Power BI as the business-facing analytics layer 

• Expansion of Data Factory and Fabric development tooling to improve orchestration and deployment 

• Ongoing enhancements to OneLake as the centralized data layer 

• Growing emphasis on governance and security through Purview 

• A bigger push toward AI agents, semantic understanding, and operational intelligence 

Rather than treating AI as a standalone feature set, Microsoft is building it directly into the broader Fabric ecosystem. That’s an important distinction—and one that makes these announcements especially meaningful for organizations trying to scale both analytics and AI responsibly. 

8 Key Announcements by Area 

1. Power BI: Smarter, More Interactive Reporting 

Power BI updates at FabCon 2026 focused heavily on making reports more interactive, more modern, and more connected to action. Microsoft continues to push Power BI beyond passive dashboards and into more translytical and workflow-enabled scenarios. 

Available now 

Several updates are already generally available, particularly in the visualization experience: 

• Translytical task flows bring new workflow capabilities directly into reports, allowing users to move from analysis to action without leaving Power BI. 

• The new button slicer and enhanced card visual improve report usability and design flexibility. 

• Modern visual defaults introduce a refreshed Fluent 2 styling experience out of the box. 

• Custom totals in tables bring more flexibility to business-facing reporting. 

• Direct Lake on OneLake remains one of the most important performance stories, reinforcing Microsoft’s push toward fast, large-scale analytics without traditional refresh bottlenecks. 

In preview 

A few Power BI features are still in preview but point to where the platform is heading: 

• Input slicer expands how users can interact with reports and data functions 

• List slicer dropdown mode improves usability for more complex filtering scenarios 

• Power BI Modeling MCP server opens the door for AI tools to work directly with semantic models 

• Mobile Copilot extends Copilot’s value to mobile users 

Coming soon 

Looking ahead, Microsoft previewed several improvements that should further streamline development and improve usability: 

• Fixed-width controls for tables and matrixes 

• Conditional formatting for lines and series 

• A new date picker mode 

• Reusable style presets for themes 

• A native Gantt chart visual 

• Additional translytical enhancements 

• TMDL View on Web and improvements to Copilot tooling/file formats 

Our take 

This was one of the clearest areas where Microsoft showed its broader vision: Power BI is evolving from a reporting layer into a more interactive decisioning layer. The combination of better visuals, semantic tooling, and translytical features suggests Microsoft wants business users to do more than consume insights—they want them to act on them. 

2. Data Factory: More Control, More Orchestration, More Enterprise Readiness 

Data Factory announcements centered on enterprise orchestration, governance, and easier operational management. Many of the updates felt geared toward making Fabric a more credible home for production-grade integration workloads. 

Available now 

Highlights included: 

• Outbound access protection for more controlled connectivity 

• Azure Key Vault support for VNet gateway connections 

• On-premises gateway auto-update to reduce maintenance overhead 

• Apache Airflow Job Fabric providers to bring Fabric operations into broader orchestration flows 

In preview 

Microsoft also previewed several features that expand flexibility and visibility: 

• SharePoint Site Picker 

• Lakehouse maintenance pipeline activities 

• Interval-based schedules 

• SSIS pipeline activity 

• Copy Job audit column 

• Workspace monitoring for Copy Job 

• Performance improvements in Copy Job 

• Data Factory MCP Server 

• Migration Assistant for Data Factory 

Coming soon 

Roadmap items included: 

• Expanded dbt support 

• Running pipelines under a service principal identity 

• Additional low-latency and CDC-oriented Copy Job enhancements 

Our take 

Data Factory is becoming more mature and more operationally complete. The emphasis on scheduling, monitoring, security, and migration support shows Microsoft understands that enterprise data teams need far more than “just pipelines”—they need repeatability, observability, and safe deployment practices. 

3. Fabric Development: A Stronger Platform for Builders 

One of the most interesting themes at FabCon was how much Microsoft is investing in the developer experience around Fabric. 

Available now 

Microsoft highlighted several capabilities already available: 

• Terraform Provider for Fabric 

• Fabric CLI 

• Workspace provisioning identity 

• Fabric local MCP 

• Quality-of-life improvements to the “stay-in-flow” development experience 

• CI/CD Library in the CLI 

In preview 

Preview features included: 

• Azure DevOps Pipelines Extension 

• Git integration enhancements 

• Variable Library deployment parameterization 

• Service principals as Fabric item identity 

• Bulk export/import through REST APIs 

• Workspace monitoring 

• Job events and alerts 

• Soft-delete and recovery 

Coming soon 

• Fabric Jumpstart, positioned as an installable accelerators and demo catalog 

Our take 

This was an important area for teams trying to operationalize Fabric at scale. Microsoft is clearly moving beyond the idea of Fabric as only a UI-first platform and is investing more seriously in automation, DevOps, and deployment tooling. That’s a strong signal for enterprise adoption. 

4. OneLake: The Center of the Fabric Story 

If AI was the headline, OneLake was the foundation underneath almost everything Microsoft talked about. Many announcements reinforced OneLake’s role as the shared storage and access layer across Fabric workloads. 

Available now 

Key available capabilities included: 

• OneLake MCP server 

• OneLake Table API 

• Shortcut transformations 

• Security UX improvements 

• Broader availability of security and storage features such as encryption and diagnostics 

In preview 

Preview items included: 

• SharePoint + OneDrive shortcuts 

• Databricks reads from OneLake 

• Security APIs for external engines, applications, and AI agents 

• Additional shortcut and mirroring sources 

Coming soon 

Roadmap highlights included: 

• Delegated OneLake shortcuts 

• Expanded Table API operations 

• File explorer enhancements like size reporting and lifecycle management 

• Eventhouse support for OneLake security 

• The GA milestone for OneLake security 

Our take 

OneLake continues to look like the backbone of Microsoft’s unified data story. The more Microsoft opens it up to external engines, APIs, and AI tooling, the more compelling it becomes as a central platform rather than simply a storage layer. 

5. Data Warehousing and SQL: Continued Convergence 

On the SQL and warehousing side, Microsoft showed continued progress toward a more unified data platform within Fabric. 

Available now 

• Mirroring for Microsoft SQL with support for more than 500 tables 

In preview 

• Database Hub 

• Database Agents 

• Multiple SQL database in Fabric enhancements 

• Data alerts and actions in Fabric Data Warehouse 

• CI/CD enabled by DacFx 

Coming soon 

• Hyperscale for Fabric SQL 

• Shortcuts from Fabric SQL 

• GeoDR 

• Additional warehouse performance improvements 

• Workspace monitoring and performance dashboard enhancements 

Our take 

This area feels like steady platform maturation. The roadmap points to a future where traditional SQL development, mirrored operational data, alerting, and CI/CD can all live more naturally within Fabric. 

6. Purview: Governance Is Becoming More Native 

Governance was another clear theme across FabCon, particularly as Microsoft tied it more directly to Fabric and AI scenarios. 

Available now 

Microsoft highlighted several generally available Purview-related capabilities: 

• Information Protection labels in Fabric 

• Data Loss Prevention policies 

• Insider Risk Management 

• Data Security Posture Management 

• Data quality and data health management in Unified Catalog 

In preview 

• Expanded restrict-access coverage for structured data workloads 

• Scoped scan support 

• Additional data health enhancements 

Coming soon 

• More restrict-access coverage for Cosmos and mirrored databases 

Our take 

This matters because organizations are not just asking whether AI works—they’re asking whether it’s governed, auditable, and secure. Microsoft’s Fabric story is increasingly pairing AI enablement with trust controls, which is exactly where many enterprises need the conversation to go. 

7. AI Agents: From Analytics to Action 

AI agents were one of the most forward-looking areas of the conference. Microsoft’s message here was that AI in Fabric is about more than chat—it’s about grounded, business-aware systems that can reason over organizational and operational context. 

Available now / currently surfaced 

• Fabric data agent fundamentals 

• Support for conversational access to lakehouses, warehouses, semantic models, KQL databases, mirrored databases, and ontologies 

In preview 

• Fabric data agents in Copilot Studio 

• Fabric data agents in Microsoft 365 Copilot 

• Data agent as MCP server 

• Operational Agents 

• The broader Fabric IQ, Foundry IQ, and Work IQ concepts 

Coming soon 

• Support for publishing to Microsoft 365 Copilot 

• Citing Fabric data agents in topics 

• Progress notification and streaming 

• Data Agent MCP Client 

• Integration with Researcher in Microsoft 365 Copilot 

Our take 

This may be the clearest picture yet of where Microsoft is headed: a world where analytics, semantic models, operational context, and AI agents work together. The interesting part is not just that agents exist—it’s that Microsoft is trying to ground them in governed enterprise data and business meaning. 

8. Cosmos DB and Translytical Patterns 

Although more architectural than product-specific, the Cosmos DB discussion helped show how Microsoft is thinking about agentic applications and operational + analytical convergence. 

Highlights included: 

• Using Fabric + Cosmos DB mirroring as an application architecture pattern 

• Leveraging Fabric User Data Functions as a custom logic layer 

• Supporting translytical workflows where analytical insights can trigger or support operational actions 

Our take 

This is one of the more compelling long-term patterns coming out of FabCon. It reinforces the idea that Fabric is not just for BI—it’s increasingly positioned as part of a broader intelligent application architecture. 

Final Takeaway 

FabCon 2026 made Microsoft’s direction clearer than ever: AI may be the headline, but data, governance, and platform integration are the real story. Across Power BI, Data Factory, OneLake, SQL, Purview, and AI agents, Microsoft is building toward a more unified environment where data can be managed, analyzed, secured, and activated in one ecosystem. 

For organizations already investing in Microsoft’s data stack, that should be encouraging. The opportunity is not just to adopt new features, but to rethink how analytics, automation, and AI can work together on top of a shared foundation. 

At Atlantic, we came away from FabCon excited about where the platform is headed—and even more interested in helping clients make sense of which of these announcements are worth acting on now versus watching as they mature. 

The post FabCon 2026: AI Gets the Headlines, But Data Is the Real Story appeared first on Atlantic | Tomorrow's Office.