Glossary · AI agentAlso called agent, agentic systemChecked 27 September 2026
What does an AI agent actually do?
An AI agent is a program that decides its own next step. It is given a goal and a set of actions it can take, and it chooses which action to run, reads the result, and chooses again until it is done or it gives up. A workflow follows a path somebody drew. An agent picks the path.
In one lineA program given a goal and a set of actions, choosing its own next step.
Why it matters to you
The distinction that matters commercially is not intelligence, it is who decides the order. A workflow is a sequence a person drew in advance, so you can read it and know what will happen. An agent selects at run time, which is the entire point and also the entire risk: the same input can take a different route on Tuesday than it did on Monday.
That makes agents right for work where the path genuinely varies, and wrong for anything irreversible. The useful design is usually both, with an agent doing the thinking and a plain deterministic worker doing anything that touches the outside world.
What it is not
A chatbot with one API call behind it is not an agent, and most things sold as agents are workflows with a model in one step.
If a proposal used this word and did not explain it, ask us.
The first conversation is with the people who would build it. If there is a fit, the Consult comes next.