Chatbot

What a Chatbot Is

A chatbot is any software that communicates through conversation — taking text input and producing text output in a back-and-forth exchange. The term covers a huge range of sophistication, from basic scripts to advanced AI assistants.

Rule-Based vs LLM-Powered

The older generation of chatbots ran on rules and decision trees. A customer support bot might recognize the phrase "cancel my order" and respond with a cancellation link. These systems could only handle scenarios their creators explicitly programmed. Anything outside those scenarios produced a confused "I'm sorry, I didn't understand that."

LLM-powered chatbots are fundamentally different. Instead of matching patterns to pre-written responses, they generate answers dynamically using a large language model. They can handle novel questions, maintain context across a conversation, and respond naturally to follow-up questions. ChatGPT, Claude, and Gemini are all examples.

What Changed With GPT-3.5 and GPT-4

Before 2022, the gap between a rule-based bot and a capable AI assistant was obvious immediately. GPT-3.5 — the model behind the original ChatGPT — crossed a threshold where the responses felt genuinely useful and natural for a huge range of tasks. GPT-4 expanded this further.

Chatbot vs AI Agent

A chatbot responds to you. An AI agent takes actions. A chatbot will tell you how to book a flight — an agent will actually book it for you by browsing the web or calling an API. The line between advanced chatbots and simple agents is blurring, but the core distinction is whether the system can act in the world or only communicate.