Skip to content
Intermediate~70 min · 6 lessons

Working With AI Agents: A Free Course

What an AI agent really is, how tools and the plan-act-check loop let it do things instead of just answering, where it fails, and how to stay in control of one.

Who this course is for

People who keep hearing that AI agents will do the work for them and want to understand the mechanism rather than the marketing. It suits anyone who has used a chat assistant and now wants to know what changes when the software can act. No programming is required, though developers will find the failure-mode lessons useful too.

What you will learn

  • What an agent actually is: a model, a set of tools, a loop, and a goal, with nothing mystical in between
  • How function calling works, and why the model never touches your files directly
  • The plan, act, check cycle that every agent run is made of, and why small per-step errors compound over long runs
  • Why prompt injection is the security problem specific to agents, and what partial defenses exist
  • How to tell which tasks suit an agent and which are better done in a chat window
  • Practical guardrails: least privilege, approval gates, hard limits, sandboxes, and logs worth reading

Before you start

None strictly, but the course assumes you have used an AI chat tool. If terms like model or context window are unfamiliar, read AI Fundamentals first.

Course contents

After this course

You will be able to look at any agent product and work out what it can actually do, what permissions it needs, and where it will break. That is the difference between delegating work sensibly and handing over access to something you do not understand.