AI AgentsLesson 1
Lesson 16 min
What is an Agent?
An LLM that can take actions in a loop, not just chat.
What you will learn
- ✓How an agent differs from a chatbot
- ✓The think-act-observe loop
- ✓When agents are (and aren't) the right tool
Explanation
A chatbot only talks. An agent is an LLM wired to tools and run in a loop: it decides an action, takes it, observes the result, and decides again — until the goal is met.
This loop (think → act → observe) lets it do real work: search the web, query a database, send an email, run code.
Agents shine for multi-step tasks with tools, but they add cost and unpredictability — for a single answer, a plain prompt is better.
Real-world use
A 'book me a flight' assistant that searches, compares, and fills a form is an agent; one that just lists tips is a chatbot.
Common mistakes
- • Building an agent for a task a single prompt could solve — it adds cost and failure modes.
Practice
Describe a task you'd give an agent and list the tools it would need.
Knowledge check
0/1 answered1. What makes an AI system an 'agent'?
Answer all questions to check.