Summary
In this module, you built the foundation you'll use to work effectively with Microsoft 365 Copilot.
Copilot works across your Microsoft 365 apps by connecting to your organizational data. Instead of returning generic responses, it grounds its output in the context of your work—your files, emails, meetings, and collaborations—while always respecting your existing permissions. When you need information beyond your organization, Copilot can also draw on the public web, allowing you to combine internal context with current external insights.
You also saw how agents extend Copilot beyond single prompts. Built-in agents like Researcher and Analyst can take on more complex, multi-step tasks, while custom agents can connect Copilot to your organization's systems and workflows.
At the same time, your data remains protected. Enterprise Data Protection commitments ensure that your prompts and responses aren't used to train Microsoft's AI models. Copilot respects sensitivity labels, follows your organization's access controls, and keeps your data within your organization's boundary. These protections are built into how Copilot works, alongside Microsoft's broader responsible AI approach.
Finally, you practiced a simple prompting framework built on four elements: context, goal, source, and expectations. This technique gives you a consistent way to guide Copilot and get more useful results. Start with a clear prompt, then refine your output through follow-up prompts.
As you continue, focus on applying these concepts in real scenarios. Use clear prompts, rely on the right data sources, and choose when to bring in agents for more complex tasks. The same patterns apply no matter which app you're working in.
Additional reading
- Write effective prompts to achieve optimal results. Deeper prompting practice including best practices, common mistakes, and the Copilot Prompt Gallery.
- Data, Privacy, and Security for Microsoft 365 Copilot. Complete details on Microsoft's data protection commitments for Copilot.