Share
What if your business users could get deep, actionable insights, not from dashboards, but by simply askin questions?
That future is no longer a vision. At #FabCon2025, Microsoft introduced one of its most powerful innovations yet: Data Agents in Microsoft Fabric. These AI-powered components act as intelligent copilots for your data, bridging natural language and structured business information to deliver insights in seconds.
What’s New and Why It Matters
Data Agents transform how we access and work with data. When someone asks a business question, the agent automatically generates the right query, whether that’s SQL for raw data, KQL for real-time events, or DAX for complex business metrics.
Let’s break down the impact.
Querying Across Diverse Data Types
Fabric Data Agents now work seamlessly across:
- Lakehouse tables – a hybrid between data lakes and warehouses, ideal for raw transactional data.
- KQL databases – real-time sources using Kusto Query Language (KQL), perfect for logs, usage data, and telemetry.
- Semantic models – curated, business-ready data models used in Power BI.
- DAX (Data Analysis Expressions) – a specialized formula language used in semantic models to define KPIs, filters, time intelligence, and other analytical logic.
Agents select the best source, or even combine multiple, to answer complex questions with precision and speed.
Ask Business Questions, Get Business Answers
Want to know simple business questions like:
“Across the industry, what are the number of games sold in each region?”
The agent might:
- Pull real-time usage from KQL logs,
- Join it with product metadata in a Lakehouse,
- Summarize it using DAX measures from a semantic model.
No code. No dashboards. Just insight.
Memory and Transparency Built In
What makes these agents truly conversational is context retention. Follow-up questions build on previous ones — filters, entities, and even specific users or products carry forward.
Each response also includes full transparency:
- How the question was interpreted,
- What query was generated,
- Which data source was used.
This builds trust while empowering iterative exploration.
Smarter With Every Use
Data Agents can be trained to align with your organization:
- Add example prompts and ideal responses,
- Include business terms and synonyms,
- Guide how specific datasets should be used.
Even if the agent stumbles, it learns from corrections, resulting in better answers over time.
Extendable Across the Microsoft AI Ecosystem
Microsoft didn’t stop at Fabric. These agents can now be:
- Published as APIs,
- Used inside Azure AI Agent Services,
- Embedded into Copilot experiences or OpenAI-based applications. 🥳
So the same data agent answering a revenue question in Fabric could be powering chatbot responses or filling out web forms in your digital workflows.
Enterprise-Ready Security and Governance
To keep all this power safe, Microsoft introduced important new features:
- Purview Information Protection Policies allow admins to control which data can be accessed and by whom—right down to individual fields.
- Integration with Insider Risk Management from Microsoft Purview helps detect and act on suspicious data access or sharing.
AI assistance doesn’t mean compromising governance, it strengthens it.
A Paradigm Shift in Analytics
This isn’t just a new feature, it’s a new way of interacting with data. Conversations are replacing dashboards. And business logic isn’t just built into reports, it’s embedded directly into how questions are asked and answered.
Now is the moment to ask:
- Are your data models ready for AI-native interaction?
- Is your business logic captured in DAX, not just PowerPoint?
- How can your teams start making smarter decisions, faster?
What’s Your Take?
Data Agents in Microsoft Fabric are more than a productivity boost, they mark the beginning of Agentic AI in enterprise environments. This shift is a game changer in how businesses interact with data: natural, contextual, and deeply integrated into decision-making processes.
Which capability excites you most, real-time analytics, self-service insights, or AI integrated directly into your workflows?
Share your thoughts or tag someone exploring Microsoft Fabric or Copilot. Let’s shape what’s next, with Xebia and Microsoft together.