AI workflow
An AI workflow is a fixed sequence of steps, some run by a model, chained together to finish a task.
What it tells you about a developer
Workflow experience means the developer has put a model into a production process with predictable behavior. That is common, valuable work, and it sits closer to ordinary backend engineering than agent work does.
A workflow follows a set path every time, the way an assembly line does. An AI workflow chains together a series of steps, where one or more of the steps calls a language model, and the order stays the same regardless of what the model decides. That is the main difference from an agent, which chooses its own steps. A developer who builds AI workflows is usually solving a repeatable process, like sorting incoming support tickets, rather than an open-ended task.