Building in the Agentic Era with Power BI and Fabric

Microsoft Build 2026 marks a major shift in how developers build data experiences with AI agents. Today we're announcing two capabilities that bring agentic analytics to the forefront: Agent Skills for Power BI, which let developers prompt an AI agent to build and refine semantic models and reports, and Fabric Apps for Semantic Models, which enable AI agents to build and deploy Fabric-native web apps on semantic models. Both capabilities will accelerate the time it takes to go from raw data to a polished analytics solution with just natural language prompts.

Agent Skills for Power BI (Preview)

In November 2025, we released our Power BI Modeling MCP, which allowed AI agents to interact with Power BI’s semantic layer, enabling massive productivity gains in semantic model authoring. Agent Skills for Power BI brings a true end-to-end agentic development workflow not just for Power BI semantic models but for Power BI reports too using Power BI Projects (PBIP). Instead of manually building the report, you can describe what you need to an AI agent in plain language, or even from a screenshot, and it will handle the development: constructing semantic models from your Fabric data, generating report pages, iterating on visuals, aligning to design best practices, and more!

 

 

The result is a fundamentally faster path from Fabric data to a production-ready report! And these Agent skills don’t just work for new reports; they work on existing solutions too, which means turning a scrappy report into a polished one is easier than ever before! Stay tuned for more details as Agent skills for Power BI will be rolling out next week. 

 

Fabric Apps on Semantic Models (Preview)

Fabric Apps introduce a new AI‑first approach to building custom web apps, with Microsoft Fabric serving as the backend and powered by the open‑source Rayfin SDK. This is going to enable a new class of rich, data experiences in Fabric!

 

In the era of AI, the cost and complexity of building applications is collapsing. Experiences that previously required deep front and backend investments can now be generated through prompt or spec-driven workflows using AI coding agents, making even consumer-grade applications far more streamlined to build. As web apps are ultimately just code, this unlocks the ability to generate nearly anything that can be described, with full control over UI, business logic, and integrations with external systems via APIs.

 

The rise of AI-generated applications extends beyond consumer scenarios. Enterprises increasingly want to build operational apps directly on governed analytics data with the same level of ease and consumer-grade polish. But for years, enterprise analytics and applications have lived in separate worlds. Semantic models provide trusted metrics and governed data, while operational apps require entirely different engineering stacks. Fabric Apps with semantic models close that gap. The same data foundation that allows analysts to build reports now makes it much easier for developers to build full-blown web applications without needing to recreate business and governance logic. With Fabric providing the shared data foundation, security, and governance for both analytics and applications, the time it takes to build and deploy operational apps on your enterprise data is significantly reduced. Scenarios like inventory management, financial planning, and staffing coordination can be developed faster and tailored more precisely to how your organization operates.