Microsoft has released Fabric Skills for GitHub Copilot, Claude, and CLI — an open-source toolkit that gives AI agents the ability to directly author and interact with Microsoft Fabric data assets, including Power BI reports and semantic models. The announcement appeared in the Power BI and Fabric June 2026 Feature Summaries, marking a meaningful step in how the data platform integrates with external AI systems.
For data professionals, this changes the tools available to them. For business leaders watching AI investment, it signals that the gap between AI assistants and production data work is narrowing faster than most expected.
What Fabric Skills Actually Does
Fabric Skills is a toolkit built by Microsoft that enables AI agents — specifically GitHub Copilot, Anthropic’s Claude, and command-line tools — to interact with Fabric environments through a defined set of capabilities.
In practical terms, this means an agent like Claude can be instructed to query Power BI semantic models, analyze Fabric data assets, and produce outputs without requiring a human to manually navigate the Fabric interface at each step. The agent works through the skill layer, which provides structured access to the platform’s capabilities in a way that external AI models can understand and act on.
Microsoft built these skills and has opened them for community contribution, meaning the set of actions available to external agents can expand as practitioners build and share new capabilities.
What This Means for Power BI Professionals
The June 2026 Fabric update also includes several capabilities beyond Fabric Skills that reflect the broader shift toward conversational and agentic analytics:
Copilot in web modeling (Preview) brings conversational AI into the semantic model authoring experience. Instead of manually editing relationships, measures, and table structures, analysts can describe what they need and have Copilot make changes directly in the browser-based modeling view.
Fabric IQ Skill provides a centralized way for end users to get answers from Power BI reports and semantic models through natural language. It is designed as the starting point for broader data discovery conversations across Fabric — with plans to extend to Fabric data agents and ontologies.
Power BI Report Authoring Agent Skills, which were announced at Build 2026, are now available as part of the broader Fabric Skills ecosystem. These allow agents to design, build, validate, and publish full reports from a description or screenshot, without manual file editing at each step.
Taken together, the direction is clear: Microsoft is building an agentic layer into Fabric that both assists human analysts and enables external AI systems to act directly on data.
What This Means for Business
For organizations already using Microsoft Fabric or Power BI, these updates are not theoretical. The ability to point Claude or GitHub Copilot at your data environment and have it produce reports, answer questions, and interact with your semantic models represents a genuine productivity shift for analytics teams.
The practical implication is that the value of a well-built semantic model just increased. When the layer above the model is increasingly agentic — AI systems querying it, authoring from it, building reports on top of it — a poorly structured model becomes a bottleneck in a way it was not before. The foundation has to be right.
For data teams, this also raises the skill bar in a specific direction. Knowing Power BI visuals and DAX is still valuable. But understanding how to build semantic models that AI agents can work with reliably — clean relationships, well-named measures, governed data assets — is becoming the skill that determines whether your AI investment works.
Enterprise DNA’s training programs cover Power BI from fundamentals through advanced model design, and our curriculum is kept current as the platform evolves. Explore Power BI and Fabric courses here — or if your business needs help building a data foundation that can support AI-driven analytics, talk to our team.