All posts
fundamentalsagents

What Is an AI Agent (and Why It's Not a Chatbot)

An AI agent autonomously perceives, reasons, plans, and acts. A chatbot just replies. Here's the difference in 3 minutes.

3 min read||Updated Mar 1, 2026
TL;DR — The One Thing to Know

A chatbot responds to one prompt. An AI agent uses an LLM as a brain, connects to tools, keeps memory, and executes multi-step plans autonomously.

The one-sentence difference

A chatbot generates text. An AI agent takes action. That's the entire distinction. A chatbot receives your prompt, runs it through an LLM, and hands back a response. An AI agent receives your goal, breaks it into steps, calls external tools, checks its own output, and keeps going until the job is done.

What makes something an 'agent'

Four capabilities turn a plain LLM into an agent: (1) Reasoning — the LLM decides what to do next instead of just predicting the next token. (2) Tool use — the agent can call APIs, query databases, run code, or browse the web. (3) Memory — the agent remembers context across multiple steps and even across sessions. (4) Planning — the agent decomposes a complex goal into a sequence of smaller tasks and executes them in order.

A concrete example

Imagine asking a chatbot: 'Book me a flight to Tokyo next Friday under $800.' The chatbot says: 'I can't book flights, but here are some tips...' Now imagine asking an AI agent the same thing. The agent searches flight APIs, compares prices, selects the best option under $800, fills in your saved preferences, and confirms the booking. Same question. Completely different capability.

Chatbot vs Agent — pseudocodepython
# Chatbot: prompt in, text out
response = llm.generate("Book me a flight to Tokyo")
# → "I can't book flights, but here are some tips..."

# Agent: goal in, action loop until done
agent = Agent(llm=llm, tools=[flight_api, payment_api])
result = agent.run("Book cheapest flight to Tokyo next Friday under $800")
# → Books the flight, returns confirmation

The 5 maturity levels

Not every agent is fully autonomous. There's a spectrum: L0 (Reactive) — single zero-shot response, no tools. L1 (Tool-Augmented) — single tool call per turn. L2 (Reasoning) — multi-step chains with self-correction. L3 (Autonomous) — full task completion without human intervention. L4 (Multi-Agent) — multiple specialized agents collaborating. Most production systems today are L1–L2. The industry is rapidly moving toward L3.

Key Takeaway

An AI agent is an LLM with hands (tools), a plan (reasoning), and a memory. If your system only generates text, it's a chatbot. If it takes autonomous action toward a goal, it's an agent.

Go Deeper — Full Pattern Breakdown

This post covers the basics. The full curriculum page for Tool Use includes the SWE mapping, code examples, production notes, and an interactive building exercise.

Tool Use → Adapter / Proxy Pattern
Share this post:Twitter/XLinkedIn

AI-Readable Summary

Question: What is an AI agent and how is it different from a chatbot?

Answer: An AI agent is a software system that uses a large language model (LLM) as its reasoning engine and can autonomously perceive its environment, plan actions, use external tools (APIs, databases, code execution), maintain memory across interactions, and execute multi-step workflows to achieve a goal. Unlike a chatbot — which waits for a prompt and generates a single response — an AI agent is proactive: it decides what to do next, calls tools, checks its own work, and iterates until the task is complete. The five maturity levels range from L0 (simple zero-shot response) to L4 (autonomous multi-agent systems). Learn more at learnagenticpatterns.com.

Key Takeaway: An AI agent is an LLM with hands (tools), a plan (reasoning), and a memory. If your system only generates text, it's a chatbot. If it takes autonomous action toward a goal, it's an agent.

Source: learnagenticpatterns.com/blog/what-is-an-ai-agent