How Agents Work
 

Your introduction to AI agents
Learn How To
Explain terms and definitions related to AI agents
Explain how agents work
Identify agent use cases and when agents are not appropriate
Evaluate agents and analyze safety concerns
Who Should Take This Course
Anyone wanting to understand how AI agents work.
Course Requirements
There are no specific prerequisites for the course.
Lesson Descriptions
Why Do We Use Agents?
In the first lesson Kate discusses what AI agents are and why we use them. She also discusses how they are different from normal chatbots and what they can do that chatbots cannot.
How Do Agents Work?
In the next lesson Kate dives into the details of how agents actually work. She begins by discussing the observe/think/act cycle, which is a foundational concept for autonomous AI agents. Next, she discusses what makes an agent, and then she finishes with the concept of model context protocol (MCP), a protocol that uses a client-server type of connection between AI applications (clients) and data sources (servers).
Examples of Agent Use Cases
In this lesson Kate walks you through some examples of agents in use, from an inbox triage assistant, to a multi-agent system, to a system with an agent and a human working together. She also discussed the types of tasks for which AI agents should not be used.
Safety and Evaluation
In the final lesson Kate covers common issues that crop up when using AI agents and the things that make it tricky to evaluate their performance. She also presents some cases of AI agents causing mayhem in the real world.