Enterprises are rapidly evolving to harness the full power of their data, and Microsoft Fabric’s latest advancements in Fabric Data Agents are setting a new standard. At Ignite 2025, Microsoft unveiled transformative updates that boost data agent capabilities, enabling organizations to extract more value from their information assets with ease and intelligence.
Key Enhancements Shaping the Future
Microsoft’s focus this year centers on two pivotal improvements:
- Expanded data reasoning: Agents now analyze both structured and unstructured data, delivering richer and more contextual insights.
- AI interoperability: New features streamline the integration of data agents with diverse AI solutions, accelerating development and adoption across the enterprise.
Unified Reasoning Across All Data Types
Fabric Data Agents have broadened their horizons. Previously limited to mirrored and SQL databases, they now natively support unstructured data, such as documents, PDFs, and media files.
This is powered by the integration of Azure AI Search Index, which lets organizations build domain-specific search indexes that connect seamlessly to data agents. The result? A unified reasoning platform that bridges structured data in OneLake with unstructured content for comprehensive enterprise knowledge discovery.
Ontology: Adding Semantics for Smart Insights
Raw data alone isn’t enough for actionable intelligence. The latest release enables agents to utilize Ontology in Fabric, blending organizational rules, semantics, and business context into reusable knowledge assets. This context-driven approach empowers data agents to deliver answers that not only make sense technically, but also align with business goals and policies.
Seamless AI Integration for Modern Workflows
Flexibility is at the core of the new Fabric Data Agents. Whether building custom copilots with Microsoft Copilot Studio or deploying multi-agent solutions through Microsoft Foundry, these agents now offer:
Hosted MCP Server Endpoints: Secure, remote connections from development environments like VS Code or other AI systems accelerate innovation without compromising on compliance or security.
Microsoft 365 Copilot Integration: Agents are now natively accessible across Microsoft 365 applications, including Teams and desktop apps. This means users can interact with enterprise data agents within familiar tools, share them securely, and collaborate seamlessly.
Security remains paramount. Every interaction with a data agent respects organizational permissions and enforces advanced controls like Row-Level Security (RLS) and Column-Level Security (CLS), ensuring analytics are both powerful and compliant.
Empowering Secure, Context-Aware Analytics
These innovations reinforce Microsoft Fabric’s vision for open, secure, and intelligent AI. By bridging disparate data sources and infusing context through ontology, Fabric Data Agents help organizations turn diverse data into actionable intelligence—empowering teams wherever they work.
Getting Started with Fabric Data Agents
Organizations are encouraged to experiment with these features by creating agents, leveraging Azure AI Search, and utilizing Ontology in Fabric. These capabilities are set to become foundational for secure and agentic analytics in the modern enterprise.
For further exploration, Ignite 2025 features sessions such as "How Fabric Data Agents are Powering the Next Wave of AI," providing deeper insights into these game-changing tools. By embracing these advancements, enterprises can future-proof their data strategy and stay ahead in the world of AI-driven intelligence.
As of the writing of this article the documentation links were still pending, check the Microsoft article for updated links.
Unlocking Deeper Data Reasoning: Fabric Data Agents Redefine Enterprise AI