Copilot Studio Overview & Architecture

• What is Microsoft Copilot Studio?
• How Copilot Studio extends Microsoft 365 Copilot
• Difference between: Out of the box Copilot, Custom Copilot agents and Power Virtual Agents (evolution)
• Copilot Studio architecture overview
• Licensing and environment requirements

Agent Design & Use Case Planning

• Identifying high value business scenarios
• Common agent patterns: Knowledge assistants, Policy & procedure agents, and Task driven workflow agents
• Writing effective agent instructions
• Guardrails, scope, and behaviour control
• Designing for trust, accuracy, and relevance

Building Your First Agent

• Creating an agent in Copilot Studio
• Agent configuration and behaviour settings
• Defining system instructions and tone
• Topics vs generative orchestration
• Managing fallback responses and clarifications
• Testing conversations in the simulator

Grounding Agents with Organisational Data

• Connecting agents to SharePoint sites & documents
• Understanding data permissions and security trimming
• Preventing hallucinations with grounding
• Best practices for structured vs unstructured data

Extending Agents with Actions & Power Platform

• Adding actions to agents
• Using Power Automate flows
• Built in vs custom connectors
• Triggering workflows from user prompts
• Error handling and user feedback

Security, Governance & Compliance

• Agent access control and sharing
• Environment strategy (Dev/Test/Prod)
• Data loss prevention (DLP)
• Responsible AI & auditability

Publishing, Deployment & Management

• Publishing agents to Microsoft 365
• Where agents appear (Copilot, Teams, web)
• Versioning and updating agents
• Usage insights and analytics