I recently had the chance to experiment with new AI-driven analytical tools, including Microsoft Fabric’s data agent. I’d like to share my experience, clarify what a data agent is, and explain how it differs from a typical AI agent.
So, here’s how I define a data agent:
A data agent is a report you can have a conversation with.
For those of us working in analytics, this could finally make two long-standing hopes come true:
#1: Analysts spend significantly less time creating visualizations.
#2: Self-service insights become more accessible to business users.
Let me go into more detail on each of these points.
Fewer Visualizations, Not Fewer Insights
I appreciate a well-crafted report that can quickly show me how the metrics I care about are performing. But with my analytics background, I also understand how reports can sometimes present metrics in misleading ways, which often leads business users to turn to analysts for KPI explanations—usually just 10 minutes before critical meetings.
This frequently traps us in a frustrating cycle: dashboards that nobody uses, and stakeholders constantly requesting “the number” on an ad hoc basis or through spreadsheets.
The good news is that visualizations and spreadsheets aren’t going away, but delivering insights now has a new approach with a Fabric data agent.
Rather than embedding queries inside charts, you can embed them within prompts and instructions, combined with the governed, consumption-ready data estate in Fabric—such as a lakehouse, warehouse, Power BI semantic models, KQL database, or even an ontology. This means the underlying data still needs to be prepared and modeled to answer business questions like “What was this week’s revenue compared to last week?“
However, from a design standpoint, instead of building a focused visual report to address this business question, you now build a focused data agent that provides this answer, along with other subsets of answers drawn from the underlying data model(s).
More specifically, the input-output flow works like this:
(1) a stakeholder asks a question, (2) the agent, powered by the Azure OpenAI Assistant API, interprets the question and determines which data source is most likely to contain the answer based on source schemas and agent instructions, (3) it generates the appropriate query (SQL, DAX, or KQL depending on the source type), (4) validates it, (5) executes it using the stakeholder’s credentials, and (6) returns the result as text or a table—not (yet) as a visual.
In short, a stakeholder’s interaction with insights through the data agent is a Q&A session built on top of a curated dataset, and drill-down visuals can be replaced with follow-up questions, such as “Can you also break the revenue down by segment?“
With this approach, it’s clear that analysts’ work no longer needs to be expressed exclusively through dashboards—the long-familiar tangible proof that the effort of capturing business logic within data models has been delivered.
Now, let’s discuss…
Self-Service Insights, Closer to Where Business Users “Live”
I mentioned earlier that reports can sometimes misrepresent metrics, but that’s not the only reason why “If you build it, they will come” rarely works for reports or analytics in general. The reality is that the knowledge barrier is often too high for people to understand the underlying semantic models and know how to use BI tools to create visuals on top of them.
While this points to data literacy—which is a change-management challenge—it’s a fact that the intended business audience, who should be report consumers, often has too much on their plate to invest time in learning BI tools for self-service analytics.
That’s why it’s important to bring insights closer to where end users “live,” which today means AI-powered tools like M365 Copilot.
With the ability to expose insights via data agents outside of Fabric, analysts can now concentrate on the analytical logic behind self-service data agents, and end users can access insights through the same AI-powered tools that support their other daily tasks—without the hassle of switching to another platform.
I should note this isn’t the only way to integrate Fabric data agents into workflows, and whether you’re a developer or a consumer, it’s helpful to understand…
The Difference Between a Data Agent and an AI Agent
So far, we’ve established that the Fabric data agent is an analytical agent focused on read-only, governed data access. It’s capable of translating natural language prompts into complex database queries that unlock insights—even outside the Fabric tenant.
On the other hand, an AI agent is defined as a system that enables Large Language Models (LLMs) to take action, not just respond to prompts, on behalf of users or other systems by accessing tools and knowledge.
In other words, the real power lies in the AI agent setup, where you can use a Fabric data agent as a specialized tool or knowledge source.
Let me illustrate this with a simple example.
Imagine an authorized user asks the AI agent to “Draft an email to the team summarizing last week’s revenue by segment.” To accomplish this task, the AI agent would, among other things, need to gather revenue insights from the enterprise database. To minimize errors in revenue calculation, the developer would design an agentic workflow that routes the input prompt to the Fabric data agent tool, which would handle the heavy lifting of determining the schema, writing the query, executing it, and returning the precise figures. Finally, the AI agent would then use those figures to complete its broader workflow and compose the email.
So what’s the difference between the two? It’s that an AI agent acts, while the data agent grounds.
Thank you for reading.
If you found this post valuable, feel free to share it with your network. 👏
Connect for more stories on Medium ✍️ and LinkedIn 🖇️.
Want to Learn More About Data Agents?
If so, check out the following resources:
Fabric data agent creation – Microsoft Fabric
Learn how to create a Fabric data agent that can answer questions about data.learn.microsoft.com
Implement Microsoft Fabric Data Agents – Training
Implement Microsoft Fabric Data Agents (chat with your data)learn.microsoft.com
