Microsoft expands chat features in Fabric to extract insights from data faster. Have a question about your datasets? Ask it in natural language.
During Build 2025, Microsoft announced new features for Fabric. The innovations aim to make it easier to extract insights from data using AI agents, manage digital twins, and integrate semi-structured data into analysis workloads. The focus is on further rolling out chat-driven data interaction and making data ‘AI-ready’ faster.
read also
Chat with your Data via Microsoft Fabric
Ask your Data
Microsoft is expanding the chat with your data experience in Fabric. Users can ask questions about various datasets and reports in natural language without first opening specific reports. This functionality is available in public preview.
References to AI agents, the main theme during Build, are not to be missed. Developers can integrate ‘data agents’ from Fabric into AI agents via Microsoft Copilot Studio. These can be deployed in Microsoft Teams and 365 apps and respond contextually to questions, respecting access rights.
Microsoft is also tapping into another AI hype: digital twins. It’s releasing a public preview of a ‘digital twin builder’ for Fabric. With this tool, organizations can build and manage virtual representations of real objects or processes. The solution works without code and integrates with OneLake and Power BI.
Preparing Data for AI
Microsoft introduces Cosmos DB in Fabric in public preview. This allows organizations to integrate semi-structured data, such as emails and documents, with existing structured and analytical data. Support for both SQL and NoSQL models should simplify building AI applications with combined data sources.
To accelerate data preparation for AI applications, Microsoft offers shortcut transformations and materialized lake views in OneLake. The so-called ‘shortcuts’ automatically apply AI-powered operations when moving data. Materialized lake views enable scheduled, automatic queries for up-to-date results.
Microsoft also announced additional features within Fabric, including:
- Real-time endpoints for ML models.
- Expansion of CI/CD support in Dataflow.
- Support for command-line interfaces and customer-managed encryption keys.
- New AI prep functionalities in Power BI.
- Preview of Translytical task flows, allowing users to perform actions from reports.
Microsoft launched Fabric two years ago. According to Microsoft’s own figures, the platform is now used by more than 20,000 organizations worldwide. The latest features are designed to help companies extract insights from their data faster and at a larger scale, with support for technical users and business decision-makers.