What is a chatbot?

Chatbot definition

A chatbot is a program designed to simulate conversation. You type (or speak) a message, it replies, and you continue from there. The term covers everything from a website widget that offers three buttons to an AI assistant that can draft a contract summary.

Old chatbots vs AI chatbots

Rule-based chatbots follow a script. They match keywords to prepared answers or walk you through a decision tree, which is why they dead-end the moment you ask something unexpected. Many customer-service bots still work this way.

AI chatbots run on a large language model. They understand free-form questions, write new text and analyze files you attach. Earlier messages stay in view because each user message and assistant response accumulates within the context window (see context window). ChatGPT, Claude, Gemini and Microsoft Copilot are the ones most professionals meet.

What it means at work

In practice the chatbot is where most people begin: you type a request, read the reply and adjust. It is a good place to learn how to give context, how to ask for a different format, and how to spot when an answer sounds right but is not. Everything else in AI at work builds on that habit.

Typical uses:

• turning meeting notes into a summary and action items;
• drafting and tightening emails;
• asking questions about a document you attach;
• explaining a spreadsheet formula or cleaning a column of data;
• brainstorming and outlining.

Who uses chatbots

Pew Research Center found that 34% of U.S. adults say they have ever used ChatGPT. In the same survey, 58% of adults under 30 say they have used it. At work, the share of employed adults who say they use ChatGPT for work has risen by 20 percentage points to 28% over two years. Pew Research Center found that, among workers who have used AI chatbots for work, the most common uses are research, editing written content and drafting written content.

How to get better answers from a chatbot

Four habits make most of the difference. State the job and the audience (“a two-paragraph update for a client who is annoyed”). Paste the material it should use instead of expecting it to know. Say the format you want, such as a table or five bullets. And treat the first reply as a draft: ask for changes, because chatbots are built for back-and-forth. These are the basics of prompt engineering, and they are why two people using the same tool get very different results.

Chatbots at work: what to watch

Data handling matters more than people expect. A personal account and a company account are different products with different terms, so check what your employer permits before pasting confidential material. See is it safe to put company data in ChatGPT.

An example

You paste a lengthy email thread into a chatbot and ask, “What did we agree on, and what is still open?” It replies in a short list. You compare it with the thread and reply to the client.

Where chatbots stop

A chatbot answers inside the chat window. It does not, by default, open your email, update a spreadsheet or run a task on a schedule. When you want that, you move to an AI agent or an automation. You can also save a good setup so you do not retype it, as with a custom GPT.

A common misconception

Chatbots are not search engines and not experts by default. They answer from patterns in their training and from what you provide, so give context, ask for sources when facts matter, and verify anything consequential. A research survey puts the failure plainly: LLMs are prone to hallucination, generating plausible yet nonfactual content. See hallucination.

Choosing between the major chatbots is covered in Claude vs ChatGPT and which AI tool to use at work. To go beyond chat and learn to use AI tools for everyday work, see Learn AI Fast, a two-day in-person class in Boston.

Sources

• Anthropic, Context windows: “each user message and assistant response accumulates within the context window”
• Pew Research Center, 34% of U.S. adults have used ChatGPT, about double the share in 2023: “34% of U.S. adults say they have ever used ChatGPT”
• Pew Research Center, 34% of U.S. adults have used ChatGPT, about double the share in 2023: “58% of adults under 30 say they have used it”
• Pew Research Center, 34% of U.S. adults have used ChatGPT: “the share of employed adults who say they use ChatGPT for work has risen by 20 percentage points to 28%”
• Pew Research Center, U.S. Workers Are More Worried Than Hopeful About Future AI Use in the Workplace: “the most common uses are research … and drafting written content”
• arXiv, Huang et al., A Survey on Hallucination in Large Language Models: “LLMs are prone to hallucination, generating plausible yet nonfactual content”