Short answer
A chatbot responds within a conversation: a message arrives, an answer goes back, and the exchange ends there. An AI agent pursues a goal across multiple steps and acts on real systems through tools, such as checking stock levels and then creating an order. The difference is whether the software only talks or also does.
Also called: Chatbot versus agent, AI assistant vs agent
There is a ladder here rather than a binary. At the bottom sits the scripted bot with fixed buttons and canned replies. Above it, a model-backed chatbot answers freely in natural language. Above that, a retrieval-backed chatbot answers from your own documents. An agent sits at the top because it is granted tools and a loop: it can look something up, decide the answer is incomplete, look again, and then change a record.
Cost, risk and build time rise at every rung, so the honest question is which rung the problem needs. A clinic that mostly answers the same nine questions about hours, location and preparation needs retrieval, not autonomy. A clinic that wants the same conversation to end with a confirmed slot in its calendar needs an agent, and needs to decide in advance what happens when the agent gets it wrong.
Common questions
Which one does a small business usually need?
Most start with a retrieval-backed chatbot, because the majority of inbound messages are repeat questions whose answers already exist in writing. Agents earn their extra cost when the valuable outcome is an action rather than an answer: a booking made, an order placed, a record updated.
Where this comes up in our work
Related terms
AI Agent
An AI agent is a system that uses a language model to decide which actions to take, then takes them through connected tools: reading a database, sending an email, updating a booking record.
Large Language Model (LLM)
A large language model is a neural network trained on large volumes of text to predict likely continuations, which lets it write, summarise, classify and translate without being programmed for each task.
Retrieval-Augmented Generation (RAG)
Retrieval-augmented generation is a technique that searches a private collection of documents for passages relevant to a question, then passes those passages to a language model as context for its answer.
Human-in-the-Loop
Human-in-the-loop is a design pattern where an AI system prepares work and a person approves, edits or rejects it before it takes effect: drafted replies queued for a one-click send, extracted invoice fields shown for confirmation.
Reading definitions because you are scoping a project? Skip ahead and just ask.