AI Agent

Chatbot vs. Agent

A chatbot responds. An agent acts.

When you ask a chatbot "book me a flight," it might give you instructions. An agent would open a browser, navigate to the airline's website, fill in your details, and complete the booking — then report back.

That distinction — responding vs. acting — is what makes something an AI agent.

The Perceive-Decide-Act Loop

Agents operate in a continuous loop:

1. Perceive — take in information from the environment (user input, tool results, memory, observations)
2. Decide — reason about what action to take next
3. Act — call a tool, write to memory, produce output, or stop

This loop repeats until the goal is achieved or the agent determines it cannot continue.

What Gives Agents Their Power

• Tools — the ability to call functions like web search, code execution, file reading, or API calls
• Memory — storing results across steps so they are not lost between actions
• Planning — breaking a complex goal into smaller sub-tasks and executing them in order

Real-World Examples

• A coding agent that reads your error message, searches the docs, writes a fix, runs the tests, and repeats until tests pass
• A research agent that searches the web, reads multiple sources, synthesizes findings, and writes a report

LLMs are the reasoning engine inside most modern agents — but the agent is the whole system, not just the model.