AI Skills in Fabric: Unlocking Insights Through Data Agents

Written by Ryan McNaught | Jun 18, 2025 4:46:15 PM

The data deluge is real, and traditional analysis methods are buckling under the weight. But what if you could harness the power of AI to transform raw data into actionable intelligence, effortlessly answering complex business questions? At FabCon in April, Microsoft Fabric unveiled a game-changer: intelligent data agents, poised to redefine how we unlock insights from our most critical data assets.

Fabric is a platform that allows developers to create and manage AI skills known as data agents. These agents are designed to operate on specific data sets, providing insights and generating responses based on user queries. Initially limited to working with Lakehouse tables, recent enhancements have expanded their capabilities to include semantic models and KQL databases, enabling them to return DAX or KQL queries for greater transparency and functionality. This empowers users with unparalleled transparency into the agent's logic and the underlying data, fostering trust and enabling deeper, custom analysis.

Understanding Data Agents

A data agent is essentially an AI skill that translates human language into precise data queries, providing answers based on the data available. This process involves:

Semantic Models: These models provide context and structure to the data, enabling agents to understand the meaning behind queries.
Lakehouse Tables: These tables store raw transactional data, offering a detailed view of operational insights.
KQL Databases: These databases contain real-time aggregated views of data, providing quick access to trends and patterns.

By integrating these elements, data agents can process complex queries across multiple data sources, delivering actionable insights with minimal error.

Enhancements in AI Skills

Recent updates to Fabric have significantly boosted the capabilities of AI skills:

Semantic Models: Enhanced support for aggregated data, allowing agents to analyze trends beyond raw transactional details.
KQL Databases: Integration with real-time view databases provides up-to-the-minute insights.
DAX and KQL Queries: Agents now return these advanced query languages, offering deeper analytical power.

These improvements not only expand the scope of questions that can be answered but also enhance the accuracy and relevance of responses.

Integration with Azure AI Agents

The integration of Fabric data agents with Microsoft's Azure AI Foundry service further enhances their utility. Utilizing an SDK, developers can interact with these agents via Open API, enabling seamless implementation in different environments.

Example Use Case: Filling Out a Form

Imagine a scenario where an agent fills out a form based on queries generated from semantic models and KQL databases:

Query Generation: The agent generates relevant questions based on provided data, such as "What is the churn rate for top players?"
Data Analysis: It retrieves detailed information about player activity, including usage hours and churn statuses.
Form Filling: Based on this analysis, it fills in details like the date of last session or churn status.

While current capabilities are still under development (e.g., speed and complexity), this example highlights the potential for agents to automate tasks that would otherwise be time-consuming and error-prone for humans.

The Future of AI Skills

The integration of semantic models with KQL databases isn't just an enhancement; it's a monumental leap forward, fundamentally changing how organizations leverage their business-specific context, from critical financial metrics (via DAX) to dynamic operational trends. This strategic move solidifies Microsoft's leadership in the GenAI race, while simultaneously elevating the importance of meticulously designed semantic models for Power BI and Fabric developers. The future of data-driven decision-making isn't coming; it's already here with Fabric's AI Skills.

From understanding the basics of how these agents operate to exploring practical applications, they offer endless possibilities for streamlining tasks and making informed decisions. As technology continues to advance, these skills will undoubtedly play an increasingly vital role in shaping the future of AI-driven decision-making across industries.

Ready to uplevel your data insights? Contact itD today and discover how to implement powerful AI Skills in your Microsoft Fabric environment, turning data into your most strategic advantage.