Consume Fabric data agent (preview)

Data agent in Microsoft Fabric transforms enterprise data into conversational Q&A systems, and it enables users to interact with their data through chat, to uncover actionable insights. One way to consume Fabric data agent is through Azure AI Agent Service, a core component of Azure AI Foundry. Through integration of Fabric data agents with Azure AI Foundry, your Azure AI agents can directly tap into the rich, structured, and semantic data available in Microsoft Fabric OneLake. This integration provides immediate access to high-quality enterprise data, and it empowers your Azure AI agents to generate actionable insights and streamline analytical workflows. Organizations can then enhance data-driven decision-making with Fabric data agent as a powerful knowledge source within their Azure AI environments.

Prerequisites

• A paid F2 or higher Fabric capacity resource
• Fabric data agent tenant settings is enabled.
• Copilot tenant switch is enabled.
• Cross-geo processing for AI is enabled.
• Cross-geo storing for AI is enabled.
• At least one of the following: A warehouse, a lakehouse, one or more Power BI semantic models, or a KQL database with data.
• Power BI semantic models via XMLA endpoints tenant switch is enabled for Power BI semantic model data sources.
• Developers and end users in Azure AI Foundry must at least have the AI Developer Role-Based Access Control (RBAC) role.

How it works

Agent Setup: In Azure AI Agent Service, create a new agent and add Fabric data agent as one of its knowledge sources. To establish this connection, you need the workspace ID and artifact ID for the Fabric data agent. The setup enables your Azure AI agent to evaluate available sources when it receives a query, ensuring that it invokes the correct tool to process the request. Currently, you can only add one Fabric data agent as a knowledge source to your Azure AI agent.

Query Processing: When a user sends a query from the Foundry playground, the Azure AI Agent Service determines whether or not Fabric data agent is the best tool for the task. If it is, the Azure AI agent:

• Uses the end user’s identity to generate secure queries over the data sources the user is permitted to access within the Fabric data agent.
• Invokes Fabric to fetch and process the data, to ensure a smooth, automated experience.
• Combines the results from Fabric data agent with its own logic to generate comprehensive responses. Identity Passthrough (On-Behalf-Of) authorization secures this flow, to ensure robust security and proper access control across enterprise data.

Adding Fabric data agent to your Azure AI Agent

You can add Fabric data agent to your Azure AI agent either programmatically or with the user interface (UI). Detailed code examples and further instructions are available in the Azure AI Agent integration documentation.

Add Fabric data agent through UI:

• Navigate to the left pane. Under Build and Customize, select Agents, as shown in the following screenshot:

This displays the list of your existing Azure AI agents. You can add Fabric to one of these agents, or you can select New Agent to create a new agent. New agent creation generates a unique agent ID, and a default name. You can change that name at any time. For more information, please visit What is Azure OpenAI in Azure AI Foundry portal.

• Initiate Adding a Knowledge Source: Select the Add button, as shown in the following screenshot:

This opens a menu of supported knowledge source types.

• Select Microsoft Fabric as the Source: From the list, choose Microsoft Fabric, as shown in the following screenshot:

With this option, your agent can access Fabric data agent.

• Create a connection: If you previously established a connection to a Fabric data agent, you can reuse that connection for your new Azure AI Agent. Otherwise, select New Connection to create a connection, as shown in this screenshot:

The Create a new Microsoft Fabric connection window opens, as shown in this screenshot:

When you set up the connection, provide the Fabric data agent workspace-id and artifact-id values as custom keys. You can find the workspace-id and artifact-id values in the published Fabric data agent endpoint. Your Fabric data agent endpoint has this format:

https://fabric.microsoft.com/groups/<workspace_id>/aiskills/<artifact-id>, and select the Is Secret checkbox

Finally, assign a name to your connection, and choose whether to make it available to all projects in Azure AI Foundry or to restrict it to the current project.

Add Fabric data agent programmatically: The following steps describe how to add Fabric data agent programmatically to your Azure AI agent in Python. For other languages (C#, JavaScript) you can refer to here.

Step 1: Create a project client

Create a client object that contains the connection string to connect to your AI project and other resources.

import os
from azure.ai.projects import AIProjectClient
from azure.identity import DefaultAzureCredential
from azure.ai.projects.models import FabricTool

Step 2: Create an Agent with the Microsoft Fabric tool enabled

To make the Fabric data agent tool available to your Azure AI agent, use a connection to initialize the tool and attach it to the agent. You can find your connection in the connected resources section of your project in the Azure AI Foundry portal.

# The Fabric connection ID can be found in the Azure AI Foundry project as a property of the Fabric tool
# Your connection ID is in the format /subscriptions/<your-subscription-id>/resourceGroups/<your-resource-group>/providers/Microsoft.MachineLearningServices/workspaces/<your-project-name>/connections/<your-fabric-connection-name>
conn_id = "your-connection-id"

# Initialize agent Fabric tool and add the connection ID
fabric = FabricTool(connection_id=conn_id)

# Create agent with the Fabric tool and process assistant run
with project_client:
    agent = project_client.agents.create_agent(
        model="gpt-4o",
        name="my-assistant",
        instructions="You are a helpful assistant",
        tools=fabric.definitions,
        headers={"x-ms-enable-preview": "true"},
    )
    print(f"Created agent, ID: {agent.id}")

Step 3: Create a thread

# Create thread for communication
thread = project_client.agents.create_thread()
print(f"Created thread, ID: {thread.id}")

# Create message to thread
# Remember to update the message with your data
message = project_client.agents.create_message(
    thread_id=thread.id,
    role="user",
    content="what is top sold product in Contoso last month?",
)
print(f"Created message, ID: {message.id}")

Step 4: Create a run and check the output

Create a run and observe that the model uses the Fabric data agent tool to provide a response to the user's question.

# Create and process agent run in thread with tools
run = project_client.agents.create_and_process_run(thread_id=thread.id, assistant_id=agent.id)
print(f"Run finished with status: {run.status}")

if run.status == "failed":
    print(f"Run failed: {run.last_error}")

# Delete the assistant when done
project_client.agents.delete_agent(agent.id)
print("Deleted agent")

# Fetch and log all messages
messages = project_client.agents.list_messages(thread_id=thread.id)
print(f"Messages: {messages}")

Related content

• Data agent concept
• Data agent scenario