Short answer
A chatbot answers from what’s written in your documentation; an AI agent also checks the account of the person asking and can take actions in your systems. So a chatbot handles questions whose answer is the same for everyone, and an agent also resolves the ones that depend on that specific customer.
In detail
Where the line is
"How do I change my plan?" is a chatbot question, because the answer is written down and it’s the same for everyone. "Why was I charged €340?" depends on that account, that month and that plan, and no article answers it, however well written. Writing more articles won’t fix that.
Why the split matters
At a software company, the second kind is usually close to half the volume, and it’s the half that takes the most time, because today it makes someone open the admin panel, check three things and come back to the conversation. A chatbot that only reads documentation can’t reach any of it.
How to tell which one you need
Take a hundred recent conversations and sort them into two piles, the ones answered with general information and the ones that need the account checked. It takes an afternoon, and that percentage decides. If the first pile is the majority, a well-built chatbot covers most of it. If the second is, a chatbot will fall short.
The word "agent" means two things
In support, an agent used to be the person answering. Since 2023 it’s also the software that answers on its own. Say "AI agent" for the second one, because both work on the same team, and a sentence like "the agent hands the conversation to the agent" confuses everyone.