Lesson 1 of 5 · about 5 min

What an agent is, and how it differs from a chatbot

When you finish, you can explain what an AI agent is and decide when a simple chat is enough.

Stuck? Ask the AI tutor

A chatbot answers; an agent acts

When you chat with an AI assistant, you ask and it answers. You then decide what to do with that answer. You copy it, edit it, send it. The assistant never leaves the chat window. This is a chatbot pattern, and it is often exactly what you need.

An agent goes further. You give it a goal, such as “find three suitable venues and draft inquiry emails,” and it works through steps by itself. It may search, read files, fill in a form, write a draft and ask you to approve. The key difference is that it takes actions, not just gives text.

A loop, not magic

Under the surface, most agents run a simple loop. Look at the goal and the situation. Decide the next step. Use a tool. Read the result. Decide again. Repeat until the goal is met or a limit is reached. The AI model does the deciding, and the tools do the doing.

There is no magic and no mind. An agent is a language model placed in a loop with tools and instructions. That is helpful to remember, because it means agents inherit the model’s weaknesses, like confident mistakes, and they can repeat those mistakes many times.

When a chat is enough

Use plain chat when the task is a single question, a draft or an explanation. “Rewrite this paragraph” does not need an agent. Using an agent for simple work adds cost and risk for no gain.

Consider an agent when a task has many steps, uses several tools, and repeats. Even then, ask what you gain. A weak plan is “Let an agent run my inbox.” A better plan is “Each morning, an agent lists new messages that mention an invoice and drafts a summary for me to read.” The second plan is narrow, checkable and low risk.

Worked example: planning a school trip

A teacher in Cairo wants to plan a museum visit. In chat, she asks for a checklist and a draft permission letter. She reads, edits and sends them herself. That is a chatbot task and it works well.

Later she imagines an agent that checks museum opening hours, compares transport options and emails the school office. She realizes the email and any booking should need her approval, and that the hours should be double-checked with the museum. She starts with the research steps only. Choosing a small first piece is the skill this course builds.

Practice now

  1. List five tasks from your week. Mark each as chat is enough or could use an agent, and explain your choice in one sentence.
  2. Pick one task and write the loop an agent would follow in plain sentences: what it looks at, what step it takes, what tool it uses and when it stops.

Use any AI assistant you have. Do not paste private, client or patient information into it.

Check yourself

1. What is the main difference between a chatbot and an agent?

2. How do most agents work under the surface?

3. Which task is best handled by a plain chat?

Ask the AI tutor

Ask anything about this lesson, in any language. The tutor is an AI: it can be wrong, and it never needs your personal details.

AI-drafted. Checked by an automated validator and a separate AI reviewer; a person has not read every lesson. Translations and tutor answers are AI-made. Found a mistake? Tell us. AI can be wrong. General information only, not professional advice. Not affiliated with any AI company named. Current University is not accredited.