Course Overview
Join our comprehensive course designed for Microsoft 365 and AI developers to master the art of developing low-code and pro-code agents and copilots for Microsoft 365. This program promises to elevate your development skills, making you proficient in the tools used to architect Copilots.
In this journey, you will first embark on an exploration of the world of Microsoft 365 and its Copilot ecosystem. You will receive an overview of the key concepts of Copilots and how to extend them in the Microsoft 365 environment.
Continuing your adventure, you will be introduced to the essence of agents, discovering the key components like knowledge and actions, and how to extend them using the Model Context Protocol (MCP). You will discern the distinctions between simple retrieval-based agents, those capable of triggering actions in existing enterprise systems, as well as autonomous or deep reasoning agents, and those orchestrated in a multi-agent environment.
Next, you will get hands-on with the practicalities of implementing low-code agents in Copilot Studio, starting with the basics. This will involve using Prompts, Topics & Variables, GenAI, Knowledge, as well as publishing your creations to different channels.
As your skills mature, you will advance to utilizing the Copilot Studio Visual Studio Code Extension, mastering automation through Agent Flows and managed Agents like the Document Processor, the integration of MCPs, Adaptive Cards, and even integrating channels like Twilio. The advanced section of Copilot Studio will conclude with the implementation of multi-agent orchestration, allowing you to create complex solutions that leverage multiple agents working together.
Your journey will then lead you to the realm of pro-code agents. Here, you will get an overview of the different implementation strategies for agents and when to use each type. You will also be introduced to the Microsoft 365 Agent Toolkit, as well as concepts of Copilot Connectors, and the new Copilot Retrieval API that allows consuming Copilot directly from your existing apps using a Microsoft Graph Endpoint.
When we get to the agents, we will start off with declarative agents, efficiently managing their capabilities like knowledge and conversation starters, API-based actions, as well as other advanced features like adaptive cards, code interpreter, and many others.
The adventure continues as you explore the creation of custom engine agents that allow you more control over aspects like choosing the right model, as well as advanced decisions such as selecting custom orchestrators like Semantic Kernel. Using the Microsoft 365 Agents SDK, you will be able to integrate Azure AI Services Agents with features like Deep Reasoning or even Multi-Agent Scenarios. You will learn to enrich these agents with custom data, leverage Azure AI Search, and employ adaptive cards. Additionally, you will implement actions based on Graph and utilize the Model Context Protocol to its fullest. We will also introduce the second flavor of custom engine agents, using the Teams AI Library V2. This will allow you to create custom engine agents that are especially suited for Teams.
Finally, you will master the critical aspects of deploying and analyzing the agents covered in this class. Ultimately, you will gain the expertise to manage agents efficiently, ensuring their seamless integration within the Microsoft 365 landscape.
At the end of this course, you will have an in-depth understanding of how to develop, deploy, and manage both low-code and pro-code agents and copilots for Microsoft 365 and always chose the right tool for the job.
Who should attend
- Microsoft 365 & AI Developers
- Decision Makers
Prerequisites
- Basic Microsoft 365 Platform Development Skills
- Basic Azure Development Skills
- C#, Typescript
Course Content
Module 1: Introduction to Copilot Extensibility & Agents
Copilot Extensibility
- Introduction to Agents and Copilots for Microsoft 365
- The Microsoft 365 platform & Copilot for Microsoft 365 Ecosystem
- Agent Types and approaches
- Extensibility Options: Adopt - Extend - Build
- Teams AI Library & Microsoft 365 Agents SDK
- Copilot & Copilot Studio Licensing Options (Full vs PAYG)
Agent Essentials
- What is an Agent?
- Key Components of an Agent: Knowledge & Tools
- Introduction to Model Context Protocol (MCP)
- Introduction & Benefits of Multi-Agent Scenarios
- Agent Builder, Copilot Studio, Microsoft 365 Agent Toolkit
Module 2: Implementing Low-Code Agents using Microsoft Copilot Studio
Copilot Studio Basics
- Microsoft Copilot extensibility fundamentals
- Introduction to Copilot Studio actions
- Implementing Agents in Copilot Studio
- Topics, Entities and Variables
- Adding Actions & Integrating Data
- Introduction to Channels
- Publish Agents to Microsoft 365
Advanced Copilot Studio Agents
- Copilot Studio Visual Studio Code Extension
- Autonomous Agents & Triggers
- Using Adaptive Cards & Code Interpreter
- Using Managed Agents: Document Processor, ...
- Use deep reasoning models for complex tasks
- Debugging Agents & Using Analytics And Insights
- Advanced Channels: Webapp, Twilio, ...
- Setup multi-agent orchestration
- Implementing multi agent-solution
Extending Copilot Studio Agents
- Prompt Actions
- Implementing & Using Connectors
- Using Agent Flows to Automate Tasks
- Use MCP in Copilot Studio
- Consuming REST APIs in Copilot Studio
Module 3: Implementing Pro-Code Agents
Overview Pro-Code Agents for Microsoft 365
- Introduction to Microsoft 365 Agent Toolkit & Agent Toolkit CLI
- Pro Code Declarative Agents vs Custom Engine Agents
- Microsoft Agent SDK vs Teams AI Library v2
- Integrate Enterprise data using Copilot Connectors
- Using the Copilot Retrieval Api
Implement Declarative Agents for Microsoft 365
- Declarative Agents for Microsoft 365 Copilot Overview
- Creating Declarative Agents using Microsoft 365 Agent Toolkit (Typescript, .NET)
- Deploying, Debugging & Fixing Errors using @m365agents and m365agentstoolkit-cli
- Agent Capabilities In Depth
- Manage Knowledge and Conversation Starters
- Implementing Key-based & EntraID Authentication
- Extend Agents Actions with API Plugins
- Enhance Presentation using Adaptive Cards
- Build Declarative Agents using TypeSpec
Implementing Custom Engine Agents using Microsoft Agent SDK
- Introduction & Benefits of using Microsoft Agent SDK
- Implementing Custom Engine Agents for Microsoft 365
- Enhance your agent using custom data and Azure AI Search (RAG)
- Implement Graph based Actions
- Using Semantic Kernel as Orchestrator
- Consuming Model Context Protocol (MCP) Tool
- Integrating Agents from Azure Agents AI Service
Implementing Custom Engine Agents using the Teams AI Library
- Introduction & Benefits of using Teams AI Library
- Dev Tunnels, Dev Tunnel CLI & Tenant preparation
- Create a custom engine agent using Teams AI Library
- Understanding Project Structure & Metadata
- Integrate Enterprise Data using RAG
- Predictable Dialogs using Adaptive Cards
- App Registrations & Single Sign-On (SSO)
- Add actions to handle complex tasks
Module 4: Deploying, Securing & Monitoring Agents for Microsoft 365
- Copilot Security & Compliance
- Introduction to Copilot Controls
- Copilot Studio Environment Architecture
- Publishing and Deploying Copilot Studio Agents
- Publish and Analyze Pro-Code Agents