Intended Audience
Professionals across operations, analysis, technical roles, business functions, and support teams who want to apply GenAI and M365 Copilot to improve productivity, decision‑making, and workflow efficiency.
Objectives
After completing this course, the learner will be able to:
■ Understand common AI terms used at work.
■ Use prompts that produce reliable, task-ready outputs.
■ Build simple, no‑code Copilot agents that automate common tasks.
■ Combine information across documents for accuracy.
■ Automate network‑related knowledge workflows.
■ Use AI safely and appropriately at work.
■ Save, share, and reuse custom-built agents across teams to enable consistent adoption.
Outline
1. Getting Started with Copilot Agents
1.1 What Copilot Agents are (and why they matter)
1.2 Designing a high-value Agent use case
1.3 Prompt engineering for Agents
Exercise: Explore pre-built Agents
Exercise: Prompt playground
2. Grounding an Agent with Your Data
2.1 Why grounding matters
2.2 Data quality and security basics
Exercise: Build a Copilot Agent
Exercise: Test queries: With 1 data source vs. without grounding
Exercise: Evaluate: Accuracy, relevance, and debug common issues
3. Building a Functional Copilot Agent
3.1 Domain-specific Multi‑File use cases
3.2 Retrieval concepts: similarity, context limits, ranking
3.3 Search criteria and prompt tuning
Exercise: Multi‑File Context Agent
Exercise: Retrieval scenarios
Exercise: Search and prompting
Exercise: Troubleshooting with Multi-File context
Exercise: Accuracy and relevance in agent response–Performance evaluation
4. Automating Work Tasks with Copilot
4.1 Workflow automation concepts
4.2 Prompt pipelining for data workflows
4.3 Ethical considerations and responsible AI practices
Exercise: Build a Process Workflow Agent
Exercise: Testing strategies and improving outputs
Exercise: Play with different Copilot Agents
Wrap-Up and Discussion
