More and more AI tools do not only answer questions. They also take actions: they search, book, send, and change things. We call these tools AI agents. In this short course, you learn how agents work and how to keep them safe and useful.
You follow an AI assistant that helps the organizer of a fictional community event, AI for Everyone. Each lesson shows real-looking examples: a goal, the steps that the agent takes, and what goes right or wrong. Then you try it yourself. We wrote the examples for teaching. A real agent may behave in a different way.
Your 25-minute route
| Chapter | What you practice | Lesson time |
|---|---|---|
| 1. What makes an agent | Follow the agent loop and tell agents apart from chatbots and workflows | 4 minutes |
| 2. How agents use tools | Separate tools that look from tools that change things | 5 minutes |
| 3. Set clear limits | Turn a vague goal into a clear goal with limits and a stop rule | 5 minutes |
| 4. Keep a human in charge | Choose which actions need approval, and write a good approval request | 4 minutes |
Plan about 18 minutes for the lessons and exercises, 4 minutes for the chapter checks, and 3 minutes for the final test. These times are estimates. Take more time if you need it.
Each chapter has one lesson and two questions. The final test has four new questions, one for each chapter. To pass a chapter check or the final test, you must answer every question correctly. After you submit, you see an explanation for each question. For multiple-choice questions, select all correct answers.
How to practice
You do not need any setup or paid tools. You can do all exercises on paper. You do not need to know how to code.
Practical Prompting is a good course to take first, but you do not need it. Remember: an agent can save you time, but you are still responsible for what it does in your name.
What you will learn
- Explain the difference between a chatbot, a workflow, and an agent
- Describe how an agent asks for tools, and who actually runs them
- Give an agent a clear goal with limits that the software enforces
- Decide which actions need human approval, and write a useful approval request
Before you start
- You know how to use everyday web applications
- You do not need programming experience, an AI account, or paid tools
Course modules
What makes an agent
Follow the agent loop, and learn how an agent is different from a chatbot and a workflow.
How agents use tools
Learn who runs an agent's tools, the difference between tools that look and tools that change things, and what to do when a tool fails.
Set clear limits
Turn a vague goal into a clear goal with allowed tools, a stop rule, and limits that the software enforces.
Keep a human in charge
Decide which actions need human approval, write an approval request that is easy to judge, and check what the agent really did.