Introduction to AI agents and prompting
Getting useful results from AI is straightforward. But getting consistently high-quality results comes down to how you use it.
Two skills make the biggest difference: knowing how to ask, and understanding which capability to use.
Copilot works best when you give it clear direction. A general prompt like "help me with sales data" gives you a broad response. But something more specific, like "analyze last quarter's sales trends and highlight the top three regional performers", guides Copilot to produce something more focused and actionable.
Copilot isn't just a single assistant. It includes different agents designed for specific tasks. For example, you might use the Researcher Agent to gather information, or the Analyst Agent to work through data in a spreadsheet. You can even create your own agents to support the way you work.
The more intentional you are with your prompts and your choice of tools, the more useful Copilot becomes in your day-to-day work.
In this module, you learn how to:
- Tap into prebuilt agents like Researcher and Analyst for specialized tasks
- Create and configure your own custom agents to automate workflows and solve problems specific to your team
- Write prompts that give Copilot the context, goals, and expectations it needs to deliver sharp, actionable results
- Use references like files, people, and topics to make your prompts more precise
By the end, you'll know how to get Copilot working smarter—not just answering questions, but anticipating what you need and delivering results that move your work forward.