Microsoft Fabric Workshop: Security and AI

Part 1 - Microsoft Fabric Security

Microsoft Fabric Security Essentials

Microsoft Fabric provides different levels of security to control access to workspaces, Fabric items and data. This module explains the Fabric permission model and how the different security levels work together.

• Control plane versus data plane security
• Overview of the Fabric permission model
• Workspace roles
• Item permissions and sharing
• Lakehouse and Warehouse access
• Shortcut security in OneLake
• LAB: Configuring Microsoft Fabric security

Fine-Grained Security with OneLake

OneLake security provides granular access control for data stored in OneLake. Security roles can be used to control access to specific tables and folders and can also restrict access at row and column level.

• OneLake Security Overview
• OneLake Security Roles
• Permissions and Role Membership
• Table and Folder Security
• Column-Level Security
• Row-Level Security
• Security across Fabric engines
• LAB: Implementing OneLake Security

Part 2 - AI Integration in Microsoft Fabric

Using Copilot in Microsoft Fabric

Copilot is available in different Microsoft Fabric workloads and can assist developers and data professionals with common tasks. This module provides an overview of the available Copilot experiences and explains where they can be useful when working with Fabric.

• How Copilot fits into Microsoft Fabric
• Licensing and prerequisites
• Using Copilot in Data Engineering
• Using Copilot with SQL and Warehouses
• Using Copilot in Data Factory
• Data privacy and security
• LAB: Using Copilot in Microsoft Fabric

Fabric Data Agents: Chat with your Data

Fabric Data Agents allow users to ask questions about data stored in Microsoft Fabric using natural language. This module explains how to create a Data Agent, connect it to Fabric data and improve the quality of the answers it provides.

• Purpose of Fabric Data Agents
• Supported Fabric data sources
• Creating and Publishing a Fabric Data Agent
• Interacting with a Fabric Data Agent
• Understanding Permission Delegation
• Improving Answers with Instructions and Example Queries

Connecting AI Agents with Fabric IQ and MCP

Fabric IQ provides business context on top of Microsoft Fabric data and makes this context available to users and AI agents. The Model Context Protocol (MCP) provides a standard way for AI applications to connect to external data and services. This module explains how Fabric IQ and MCP can be used to make Fabric data and capabilities available to AI agents.

• Introduction to Fabric IQ
• Business context and intelligence in Fabric
• Introduction to the Model Context Protocol (MCP)
• Microsoft Fabric MCP Servers
• Using a Fabric Data Agent as an MCP Server
• Connecting External AI Agents to Microsoft Fabric
• Authentication, Permissions and Security
• LAB: Connecting an AI Agent to Microsoft Fabric using MCP