AI for Developers – Agents in Copilot/VS Code (MS-4010 / MS-4015)

AI for Developers – Agents in Copilot/VS Code

This course provides a comprehensive introduction to extending Microsoft 365 Copilot using declarative agents, API plugins, Copilot connectors, and custom engine agents.

It focuses on building declarative agents using Visual Studio Code and the Microsoft 365 Agents Toolkit to extend Microsoft 365 Copilot. Participants learn to create API plugins, implement Adaptive Cards for rich responses, manage authentication, and build Copilot connectors.

It covers building custom engine agents using the Microsoft 365 Agents SDK, leveraging Azure OpenAI, and implementing Retrieval Augmented Generation (RAG) with enterprise data.

Audience

• Developers and IT professionals expanding skills in Microsoft 365 development and automation
• Developers interested in creating custom agents using the Microsoft 365 Agents SDK

Prerequisites

Declarative Agents

• Familiarity with Microsoft 365, Microsoft 365 Copilot, and Microsoft Teams
• Familiarity with Visual Studio Code and the Microsoft 365 Agents Toolkit
• Familiarity with TypeScript, JavaScript, .NET, and/or C#

Custom Engine Agents

• Familiarity with Microsoft 365 Agents Toolkit, Visual Studio Code, Microsoft Teams, Microsoft 365 Copilot, and creating chat bots
• Proficient understanding of Azure Resources such as Azure OpenAI
• Proficient understanding of Microsoft 365 Agents Toolkit and Visual Studio Code

Course Requirements

Students must bring their own Azure Subscription with Azure OpenAI resources enabled and a Microsoft 365 Developer Account to complete labs.

By the end of this course, learners will be able to:

• Gain knowledge of declarative agents and their role in Microsoft 365 Copilot
• Create Adaptive Cards to enhance user interactions
• Implement action APIs and integrate Microsoft Graph connectors
• Build declarative agents using Visual Studio Code and the Microsoft 365 Agents Toolkit
• Build Microsoft 365 Copilot connectors using the Microsoft 365 Agents Toolkit
• Classify the uses of the Microsoft 365 Agents SDK suitable for business needs
• Deploy an app using Azure OpenAI resources to connect your data
• Understand and implement Retrieval Augmented Generation (RAG) with Azure OpenAI Service
• Build custom engine agents using Microsoft 365 Agents SDK

AI for Developers – Agents in Copilot/VS Code Outline

Introduction to Declarative Agents for Microsoft 365 Copilot

• What declarative agents are
• How declarative agents work
• When to use declarative agents
• Lab: Build Your First Declarative Agent Using TypeSpec
  • Scaffold a new declarative agent
  • Write instructions and conversation starters
  • Integrate knowledge and actions

Introduction to Actions with API Plugins for Declarative Agents

• What API plugins are
• How API plugins work
• When to use API plugins
• Lab: Build Your First Action for Declarative Agents with API Plugin
  • Create a declarative agent with an API plugin
  • Build an API plugin using Visual Studio Code
• Lab: Use Adaptive Cards to Show Data in API Plugins
  • Return rich responses from API plugins
  • Use Adaptive Cards to present structured data
• Lab: Authenticate Your API Plugin for Declarative Agents with Secured APIs
  • Authentication concepts for API plugins
  • Integrate APIs secured with keys
  • Integrate APIs secured with OAuth

Introduction to Microsoft 365 Copilot Connectors

• External connections and schemas
• Data retrieval and transformation
• Access controls and data security

Build Your First Microsoft 365 Copilot Connector

• Create a connector project in Visual Studio Code
• Connect to external data sources
• Run a connector and ingest data

Monitor and Maintain Microsoft 365 Copilot Connectors

• Monitoring connector operations
• Troubleshooting errors
• Managing permissions and security

Understand Microsoft 365 Agents SDK

• Leverage the Copilot stack to accelerate your AI build
• Explore Microsoft 365 Agents SDK
• Classify uses suitable for business needs

Utilize an Azure OpenAI Model to Create an Agent

• Setting up the development environment
• Integrating Azure OpenAI and other AI services
• Understand Retrieval Augmented Generation (RAG) with Azure OpenAI Service
• Exercise: Implement RAG with your own data and Azure OpenAI models
• Exercise: Install and set up Microsoft 365 Toolkit for Visual Studio Code

Build Your Custom Agent Using Microsoft 365 Agents SDK

• Create a new agent project with the Microsoft 365 Agents SDK
• Exercise: Create your custom engine agent
• Implement the agent’s logic
• Exercise: Configure your custom engine agent
• Deploy the agent to Azure App Service and Teams
• Test and troubleshoot the agent locally and in Teams

Course Deliverables

Upon completion, participants will have built:

• A declarative agent project
• API plugin implementation with Adaptive Cards
• A Copilot connector project
• A custom engine agent configured for Microsoft Teams