Build your first AI agent in Copilot Studio—free and hands-on. Amelia Roberts shows how to design goals, connect actions and data, test conversations, add guardrails, and publish safely to channels like Microsoft Teams.
Build a working AI agent in a single day. In this free, beginner-friendly workshop, Amelia Roberts walks you through the essentials of Copilot Studio—planning what your agent should do, grounding it with the right data, and shaping conversations that stay helpful and on-task. You’ll explore the building blocks step by step: prompts, actions/tools, connectors, and simple guardrails that keep experiences safe and reliable.
From first sketch to first publish, you’ll practice how to test, iterate, and improve your agent’s responses. We’ll also touch on hand-offs to humans, basic authentication, and options for sharing your agent in channels like Microsoft Teams or the web. By the end, you’ll have a clear recipe you can repeat at work—design, wire up, test, and ship agents that actually get things done.
Course Outline ( Free Preview)
Module 01 - Introduction
Amelia welcomes you to the workshop, outlines the goals for the day, and shows how Copilot Studio helps you design agents that actually complete tasks—not just chat. You’ll see the agenda, what you’ll build, and how each module stacks to take you from idea to a published, testable agent.
Module 02 - Copilot Chat and Agents
Learn the difference between conversational Copilot experiences and purpose-built agents. Amelia explains when simple chat is enough, when a dedicated agent makes sense, and how grounding, memory, and actions shape real outcomes for users across channels like Teams and the web.
Module 03 - Building in Agent Builder
Get oriented to Agent Builder, the place where you create, configure, and manage your agent. You’ll tour key panels, learn how versions and publishing work, and see how to keep environments tidy so you can iterate quickly without losing track of what changed.
Module 04 - Copilot Studio
Step back and view the whole Copilot Studio landscape. Amelia connects the dots between data connections, tools/actions, security, analytics, and deployment. You’ll learn where to configure each piece and how governance fits into day-to-day agent building.
Module 05 - Planning Your Agent
Before writing a single instruction, you’ll define the agent’s mission, users, success criteria, and guardrails. Amelia walks you through scoping tasks, mapping happy/edge paths, and identifying required data and tools so your build has a clear blueprint.
Module 06 - Creating Your Agent
Time to build. You’ll create a new agent, set up basic properties, connect initial data, and run your first smoke tests. Amelia shares a simple checklist to verify the agent responds, follows its role, and is ready for deeper authoring.
Module 07 - The Authoring Canvas and Leveraging Tools
Explore the authoring canvas where logic comes together. You’ll add and configure tools (actions, connectors, APIs), pass parameters, and handle responses. Amelia covers patterns for safely calling external systems and returning user-friendly answers.
Module 08 - Agent Flows
Design step-by-step flows that guide how your agent thinks and acts. You’ll create branches for different scenarios, add validations, and loop where needed—while keeping flows readable. Expect tips for error handling, fallbacks, and human handoff.
Module 09 - Building Autonomous Agents with Copilot Studio
Introduce autonomy thoughtfully. Amelia shows how to let agents plan sub-tasks, call tools, and self-correct within clear limits. You’ll add light guardrails, success checks, and escalation rules so independence improves outcomes—not risk.
Module 10 - Crafting Agent Instructions
Strong instructions are your agent’s operating manual. Learn to write crisp roles, capabilities, constraints, and examples that reduce ambiguity. You’ll practice iterative prompting, add retrieval hints, and bake in tone and compliance requirements.
Module 11 - Fundamentals of Autonomous Agents: Tools
Deep-dive on tools for autonomous behavior: selecting the right action, shaping inputs/outputs, and chaining calls safely. Amelia shares patterns for idempotency, timeouts, and retries so complex sequences remain reliable.
Module 12 - Fundamentals of Autonomous Agents: Agent Flows
Extend your flows for autonomy: plan-execute-reflect loops, checkpoints, and recovery paths. You’ll learn how to log decisions, preserve context, and keep flows maintainable as capabilities grow—without creating brittle logic.
Module 13 - Class Wrap Up
Close with a practical launch checklist: test plans, analytics to monitor, access and privacy settings, and publishing steps. Amelia recaps key patterns, suggests next experiments, and points you to resources to keep improving your agent after day one.
Amelia Roberts has a Master's in Education and spent 10 years teaching grades K-12 before joining the Pragmatic Works Training Team. Her goal at Pragmatic Works is to provide energetic and engaging training that builds up your confidence in various programs. Her outside hobbies include coaching a dance team, reading, and experiencing life with her son.