Chatbot
A chatbot is a software program that holds a conversation with people in everyday language, through text or voice, answering questions and helping with tasks as if you were messaging a person.
What is Chatbot?
A chatbot is, at its simplest, software you talk to. Instead of clicking through menus or filling in forms, you type or speak a request in ordinary language and the chatbot responds in kind. The idea is not new — the earliest chatbots date back to the 1960s — but for most of that history they were far more limited than the name suggested. Early chatbots, and most of the ones businesses used until recently, were essentially scripted: they scanned your message for a few keywords and replied with pre-written answers, following a fixed decision tree. They worked when you asked exactly what they expected and fell apart the moment you didn't, which is why so many older chatbots were quietly infuriating to use.
What changed everything was the arrival of large language models. Modern chatbots built on them are a different kind of thing entirely. Rather than matching keywords to canned responses, they actually interpret what you wrote — including messy phrasing, follow-up questions, and requests no one scripted in advance — and generate a fresh, relevant reply on the spot. This is the leap that took chatbots from a frustrating novelty to genuinely useful tools, and it is why the word can be confusing today: a "chatbot" might mean a primitive keyword-matcher on an old website or a sophisticated AI assistant that can draft your emails and explain your tax return. Same label, vastly different capability.
It helps to keep one distinction in mind: the chatbot is the interface, not the intelligence. It is the conversational front end — the chat window, the messaging experience — while the actual understanding usually comes from an underlying model doing the heavy lifting behind it. The same powerful language model can sit behind a customer-service chatbot on a retailer's site, a coding helper for developers, and a general assistant on your phone; what makes each one a "chatbot" is simply that you interact with it by chatting. That framing also explains why quality varies so wildly: a chatbot is only as capable as the system behind it and the care taken in setting it up.
Real-world example
A family is planning a weekend outing and someone lands on a science museum's website late at night, long after the ticket office has closed. Instead of hunting through pages, they type into the little chat bubble in the corner: "Coming Saturday with two kids under 10 — how much for tickets, and is there parking nearby?" A good modern chatbot reads that whole messy, multi-part question, answers both halves — the family ticket price and the parking situation — and might add that under-5s are free. An older keyword chatbot would likely have caught only "tickets," dumped a generic pricing link, and missed the parking question entirely. The difference between those two experiences is the difference between chatbots then and now.
Related terms
Frequently asked questions
What is the difference between a chatbot and ChatGPT?
ChatGPT is a specific, very capable chatbot (and AI assistant) made by OpenAI; "chatbot" is the general category it belongs to. All chatbots are programs you converse with, but they range enormously in ability — from a basic keyword-matcher on a company website to an advanced system like ChatGPT built on a large language model. So ChatGPT is a chatbot, but most chatbots are not nearly as capable as ChatGPT.
Are all chatbots powered by AI?
No, and this trips people up. Plenty of older or simpler chatbots are just rule-based scripts: they look for keywords and return pre-written replies, with no real understanding and no AI in any meaningful sense. Modern chatbots built on large language models genuinely interpret what you say and generate original responses. Both wear the same "chatbot" label, which is why two of them can feel worlds apart — one rigid and easily stumped, the other fluent and flexible.
What can chatbots be used for?
A great deal, depending on how capable they are. Common uses include answering customer questions and handling support, guiding people through a website or signup, booking and scheduling, and acting as a personal assistant that drafts text, explains things, or talks you through a problem. The basic appeal is always the same: let people get what they need by simply asking in plain language, at any hour, instead of navigating menus or waiting for a human. How well that works comes down to the strength of the system behind the chat window.