Updated July 2026

Microsoft Copilot Studio

Class Duration

14 hours of live training delivered over 2 days.

Student Prerequisites

Familiarity with Microsoft 365 apps such as Teams, SharePoint, and Outlook
Comfort with low-code tools and everyday business software
No professional programming experience required

Target Audience

Business analysts, makers, power users, and team leads who want to build custom AI agents in Microsoft Copilot Studio. Ideal for organizations adopting Copilot Studio to automate conversations and processes, ground agents in their own knowledge, and publish agents into Teams and Microsoft 365 Copilot. A deep technical background is helpful but not required, and the course builds every concept from the ground up.

Description

Microsoft Copilot Studio is a low-code environment for building, testing, and publishing AI agents that answer questions and take action on your behalf. This hands-on course takes participants from their first agent to a published, dependable assistant grounded in trusted organizational knowledge. Participants learn the anatomy of a Copilot Studio agent, how generative answers work and how to ground them in SharePoint, files, websites, and Dataverse, and how to shape conversations with topics, trigger phrases, questions, entities, and variables. The course then adds action: connectors and Power Automate flows that let an agent do more than answer, along with deterministic workflows used as tools. Participants practice testing and tracing agent behavior, reviewing analytics, and publishing to channels such as Teams, a website, and Microsoft 365 Copilot. Security and governance are covered throughout so agents are shared safely and stay within organizational policy. Every topic is practiced against realistic workplace scenarios.

Learning Outcomes

Explain what a Copilot Studio agent is and how its instructions, topics, knowledge, and tools fit together.
Build an agent from a description and refine its behavior in the maker experience.
Ground an agent in knowledge from SharePoint, files, websites, and Dataverse.
Use generative orchestration and authored topics to produce reliable responses.
Design conversations with trigger phrases, questions, entities, variables, and branching.
Add actions, connectors, and Power Automate flows so an agent can complete tasks.
Test, trace, and analyze an agent, then publish it to Teams, the web, and Microsoft 365 Copilot.
Apply authentication, sharing, and governance controls to keep agents safe and compliant.

Training Materials

Comprehensive courseware is distributed online at the start of class. All students receive a downloadable MP4 recording of the training.

Software Requirements

Students need access to Microsoft Copilot Studio in a Power Platform environment, a Microsoft 365 account with Teams and SharePoint, and a web browser. Access to Power Automate is recommended for the automation topics. No local installation is required.

Training Topics

Getting Started with Copilot Studio

What Copilot Studio is and where agents run
The maker experience and the current agent experience
Environments, licensing, and the Copilot Studio home
Anatomy of an agent: instructions, topics, knowledge, and tools
Building a first agent from a plain-language description

Generative Answers and Knowledge

How generative answers ground responses in your content
Connecting SharePoint, files, websites, and Dataverse
Generative orchestration compared to authored topics
Citations, accuracy, and reducing made-up answers
Content moderation and response quality

Designing Conversations with Topics

Topics as focused, reusable conversations
Trigger phrases and trigger types
The authoring canvas: messages, questions, and nodes
System topics, fallback, and escalation
Trying topics in the test pane

Variables, Entities, and Logic

Capturing user input with questions
Prebuilt and custom entities
Variables and their scope across a conversation
Conditions, branching, and looping
Shaping values with Power Fx formulas

Actions and Automation

Adding actions so an agent can do more than answer
Connectors and prebuilt actions
Power Automate flows as agent tools
Passing inputs and handling outputs and errors
Deterministic workflows for predictable behavior

Enhancing Agents with AI

Generative orchestration of tools and knowledge
Writing instructions that produce reliable behavior
Adding and tuning AI prompts
Balancing generative and authored responses
Iterating toward consistent results

Testing and Publishing

The test pane and conversation tracing
Reviewing analytics and agent performance
Publishing an agent
Channels: Teams, custom websites, and Microsoft 365 Copilot
Environments and the basics of application lifecycle management

Security and Governance Basics

Authentication options for agents and users
Data loss prevention and environment policies
Sharing agents and assigning roles
Keeping agents within organizational policy
Responsible AI considerations for everyday agents