At its FabCon and SQLCon Barcelona 2026 event, Microsoft laid out what it is calling the next chapter for Power BI. Two of the announcements matter most if your team builds or reads Power BI reports today: a new way to turn a semantic model into a working app without writing code, and a path for that same semantic model to show up directly inside Microsoft 365 Copilot. Here is what changed, and why it is worth a look before your next planning cycle.
Describe the app you need, and Power BI builds it
Power BI Desktop is getting a new, agentic app building experience, powered by Fabric Apps. Microsoft describes it as starting "with a trusted semantic model," where users "describe the application they need, generate it using natural language, preview and edit it in real time, and publish it to Microsoft Fabric without starting from a blank canvas."
That is a different starting point than most low-code app projects. Instead of a developer building screens against a data source from scratch, the app comes out of a semantic model your team has already built and trusted, the same measures and relationships that already drive your reports.
Microsoft is also widening who can use it. The new experience will not be limited to customers on Fabric capacity: according to Microsoft, "apps in Power BI will be available in preview in the coming weeks to Power BI Pro and Power BI Premium Per User customers as well as Microsoft Fabric capacity customers." If your organization is still on Pro or Premium Per User licensing rather than full Fabric capacity, this is one to watch for in the coming weeks rather than something to assume is Fabric-only.
The semantic layer keeps expanding
The second change is about how much weight Power BI's semantic layer already carries, and where it is headed next. In its post "Power BI's next chapter: The evolution of business intelligence", Microsoft calls the semantic layer "an organization's most strategic asset," and backs that up with scale: "More than 22 million Power BI semantic models are in use every day, making Power BI the world's most widely adopted semantic layer." Microsoft adds that the semantic layer is "increasingly" also what "grounds agents, applications, and Microsoft Copilot in that same business meaning."
At FabCon Barcelona, Microsoft gave "an early look at semantic views," describing them as "allowing users to define governed business metrics and semantics directly where data lives in OneLake and reuse them across and beyond Fabric." That is a preview, not a generally available feature, but the direction is clear: Microsoft says its goal is "to bring semantics closer to where the data lives while making it open, reusable, and interoperable."
For a team already depending on Power BI for daily reporting, this is the piece that ties the other two together. The same semantic model that defines what a "customer" or a "completed order" means for your reports is becoming the shared business meaning that apps, agents and Copilot draw on too.
The data model behind your reports is reaching Copilot
The second shift is about where your semantic model can be used, not just how reports get built. Microsoft's own description is direct about this: "Business definitions, key metrics, and organizational knowledge captured in trusted Power BI semantic models and reports can now directly inform decisions and actions in Microsoft Copilot." That connection runs through Fabric IQ, which Microsoft says combines "unified data from Microsoft OneLake, trusted metrics from Power BI semantic models, and operational context from ontologies and real-time intelligence."
In practice, that means two things are now generally available rather than in preview:
- Fabric IQ in Microsoft Copilot Chat and Copilot Cowork, so a question asked in Copilot can be answered using the same governed semantic model behind your Power BI reports, not a generic answer from the web.
- The Fabric IQ MCP server, which Microsoft's documentation confirms "is generally available." It is "a remote Model Context Protocol (MCP) server that connects AI agents, applications, and MCP-compatible clients to Power BI reports and semantic models in Microsoft Fabric," exposing "a read-only toolset that helps an agent find Power BI content, inspect report and semantic model metadata, locate relevant values, and run DAX queries against data the signed-in user can access."
That second piece is the one worth pausing on. It means tools outside Power BI itself, including an AI agent built on a different platform, can now be pointed at your semantic model and ask it questions in DAX, within whatever access the signed-in user already has. Microsoft is explicit that this does not bypass existing security: the server "applies the caller's existing Microsoft Fabric permissions," and row-level and object-level security "continue to restrict returned data."
What it means for your team
Put together, these two changes point in the same direction. The semantic model behind your Power BI reports is becoming something an agent, whether that is Copilot or another tool connected through MCP, can reason over directly. It is no longer only something an analyst opens and queries by hand.
That raises the value of getting the semantic model right in the first place. A model with clean measures, clear relationships, and well-managed row-level security now feeds more than just the reports your team looks at each week. It is also the foundation for the apps and AI answers your organization will build on top of it next.
Where to start
If your Power BI semantic models were built a few renewals ago, or grew without much governance, this is a good moment to take stock before deciding how far to extend them into apps or Copilot. See how we approach this kind of work on our services page, or tell us where your reporting is falling short today.
