What Participants Will Need (Bring-Your-Own-Device)

Prerequisites

• • Basic understanding of Microsoft Copilot (Copilot Level 1 or equivalent hands-on experience).
  • • A modern web browser — Microsoft Edge (recommended) or Google Chrome, kept up to date, for Copilot Chat, Cowork, Copilot Studio, and Power Platform Admin Center
    • Microsoft 365 apps (desktop or web) — Outlook, Word, Excel, PowerPoint, Teams, and SharePoint/OneDrive, since exercises pull real context from these apps
    • The Microsoft 365 Copilot app (desktop or browser-based), which is where Cowork is launched from
    • Stable internet access sufficient for cloud app use and video throughout the day
    Not required, but useful: the Teams mobile app for testing published agents on a second device.
    

Target Audience for this Course

• • Business users
  • Power users
    
  • Citizen developers
  • Anyone who want to move beyond everyday Copilot Chat use into building repeatable workflows and custom agents.

Course Content

• Module 1 — Foundations: Workflows, Agents & Agentic AI

  This module sets the conceptual groundwork for the rest of the day, establishing shared vocabulary before participants start building.

• Basic workflows and agents in Microsoft 365

  An overview of where workflows and agents already show up across Microsoft 365 today — from Power Automate flows to built-in Copilot agents in apps like Outlook, Teams, and SharePoint. Participants see concrete, familiar examples before the terminology gets more advanced.

• Introduction to Agentic AI

  A plain-language explanation of what makes an AI system “agentic” — the ability to reason over a goal, choose actions, use tools, and adapt when the situation changes, rather than simply responding to a single prompt. Includes a short discussion of where agentic AI adds value versus where a simple workflow is more appropriate.

• Copilot Chat vs Cowork vs Agents

  A side-by-side comparison of the three main ways people interact with Copilot: single-turn conversational Copilot Chat, the multi-step task delegation of Cowork, and purpose-built custom agents created in Copilot Studio. Participants leave with a simple decision framework for choosing the right tool for a given task.

• Understanding Work IQ

  An introduction to Work IQ, the layer that lets Copilot understand organisational context — relationships between people, files, meetings, and projects — so that responses and automations are grounded in real, relevant business data rather than generic answers.

• How Cowork uses Microsoft Graph

  A look under the hood at how Cowork uses Microsoft Graph to securely access emails, files, calendars, chats, and other Microsoft 365 data on the user's behalf — always within existing permissions — to complete multi-step tasks.

• Module 2 — Getting Started with Copilot Cowork

  A hands-on introduction to Cowork, covering setup, licensing, and how to write requests that reliably produce the outcome you want.

• Getting Started with Cowork

  A guided walkthrough of the Cowork interface: starting a new task, monitoring progress, reviewing intermediate steps, and receiving the final output. Participants complete their first simple Cowork task during this session.

• Access and licensing

  An overview of the licensing requirements for Cowork, how access is provisioned by IT administrators, and what to check if a feature appears unavailable.

• Writing Effective Cowork Requests

  Practical techniques for writing requests that give Cowork enough context and structure to succeed on the first attempt — covering scope, source data, format expectations, and how to iterate when the first result isn't quite right.

• Outcomes vs prompts

  A key mindset shift for working with Cowork: describing the outcome you want (e.g. “prepare my board pack from this month's project updates”) rather than scripting the exact steps, and letting Cowork plan and execute the task.

• Context and approvals

  How Cowork gathers context from relevant files, emails, and conversations, and where human approval checkpoints sit in the process — particularly before any action that sends, shares, or modifies data on the user's behalf.

• Module 3 — Cowork in Practice

  Moving from mechanics to application: realistic use cases, recurring automation, reusable building blocks, and the governance guardrails that keep Cowork safe to use at scale.

• Practical Use Cases

  A set of realistic, role-based examples — such as summarising a week's meetings, drafting a status report from multiple sources, or preparing onboarding materials — that participants can adapt to their own work. Time is set aside for participants to try a use case relevant to their own role.

• Scheduled tasks

  How to configure Cowork tasks to run automatically on a recurring schedule (daily, weekly, or at a set trigger point), turning a one-off request into a standing automation.

• Reusable skills

  An introduction to packaging a well-tested request as a reusable “skill” that can be triggered again with new inputs, reducing the need to redraft the same instructions from scratch each time.

• Governance & Security

  An overview of the guardrails around Cowork: data boundaries, permission inheritance, audit visibility, and administrator controls — plus practical guidance on what sensitive data should and shouldn't be routed through automated tasks.

• Module 4 — Building Agents with Copilot Studio

  The second half of the day shifts from using Microsoft-built automation (Cowork) to designing custom agents purpose-built for a specific process, department, or audience.

• Introduction to Copilot Studio

  An overview of Copilot Studio as a low-code platform for designing, testing, and publishing custom agents, and how it complements Cowork and out-of-the-box Copilot experiences.

• Creating a new agent in Copilot Studio

  A guided, hands-on walkthrough of creating a new agent from scratch: naming and describing the agent, setting its instructions and tone, and configuring its initial knowledge sources.

• Working with environments

  How environments are used to separate development, testing, and production agents, and why this separation matters for safe iteration and change management.

• Adding topics

  How to define topics — specific conversation scenarios with trigger phrases and conversational flows — so the agent responds appropriately to common questions and requests.

• Taking action with tools

  How to connect an agent to tools and connectors so it can do more than converse — retrieving live data, updating records, or calling external systems and APIs as part of a conversation.

• Automation with triggers

  How agents can be started automatically by an event — such as a new item in a list, an incoming email, or a scheduled time — rather than waiting for a user to start a conversation.

• Agents as assistants

  A look at how agents can be embedded as ongoing assistants within Microsoft 365 apps and channels (Teams, Microsoft 365 Copilot, websites), supporting employees or customers directly where they already work.

• Testing your agents

  How to use Copilot Studio's built-in test pane to validate topics, trigger phrases, and tool calls before publishing — and how to iterate based on test conversations.

• Publishing agents and analysing performance

  The final step of the agent lifecycle: publishing an agent to its intended channel(s) and using built-in analytics to monitor usage, resolution rates, and areas for improvement over time.

• Course Wrap-Up

  The day closes with a short recap of when to reach for a workflow, when to reach for Cowork, and when a custom Copilot Studio agent is the right tool — along with a discussion of next steps for applying these skills back in participants' own roles.