Agent
A program that takes a goal, works out the steps itself, and uses tools to carry them out, instead of answering one question and stopping.
The difference from a chatbot is the loop. A model answers; an agent reads the result of what it just did, decides what follows, and keeps going until the goal is met or it runs out of room. Memory, tools, and permissions all exist to serve that loop.
The loop is also where the risk lives. An agent that can read a page can be instructed by that page, and an agent that can run a command can run the wrong one. That is why permission modes, a confirmation before every step that cannot be undone, and a readable trail of what happened matter more here than they do for a single answer.
For someone who builds sites, an agent is also a visitor. It reads a page as structure rather than as a picture, so semantic HTML, real labels on fields, and predictable forms decide whether it can use the site at all.