Building Agents in Microsoft 365 Copilot training course

• Writing effective instructions and prompts
• Structuring agent behaviour and tone
• Defining purpose and scope (use-case driven design)
• Connecting agents to Microsoft 365 data (Graph-based knowledge)
• Reusable agent patterns (FAQ bot, document assistant, meeting assistant)

Hands-on lab:

• Improve and refine your agent’s responses
• Add contextual instructions and constraints

Outcome:
Design agents that produce consistent, useful outputs

Module 4: Working with Knowledge & Data

• Using Microsoft 365 content (SharePoint, Teams, emails)
• Grounding agents in organisational knowledge
• Data security and permissions (respecting M365 boundaries)
• Best practices for accurate responses

Hands-on lab:

• Build an agent that answers questions from internal documents

Outcome:
Create knowledge-driven agents for real business scenarios

 Module 5: Use Cases & Business Scenarios

• Common agent scenarios:
  • Document assistant
  • Meeting follow-up agent
  • Research assistant
  • Task and productivity assistant
• Embedding agents into daily workflows
• When to move to Copilot Studio (decision guidance)

Hands-on lab:

• Build a scenario-based agent (e.g. onboarding assistant or helpdesk FAQ)

Outcome:
Apply agents to real workplace use cases

Module 6: Testing, Sharing & Governance

• Testing and refining agent behaviour
• Sharing agents with teams
• Managing versions and updates
• Governance and responsible AI considerations
• Limitations of Microsoft 365 Copilot agents

Hands-on lab:

• Test and publish your agent

Module 7: Wrap-Up & Next Steps

• Recap of key concepts
• Agent design checklist
• Scaling beyond M365 (intro to Copilot Studio)
• Further learning resources

 Practical Outcomes

By the end of the course, participants will:

• Build multiple working Copilot agents
• Use natural language to define agent behaviour
• Connect agents to Microsoft 365 knowledge
• Apply agents to real business scenarios