Deciding whether a task is ready for an AI workflow is often harder than building the workflow itself. Many tasks get labeled "not a good fit" when the real issue is that they haven't been structured with automation in mind. This session covers two things worth knowing before building anything: how to assess whether a task is genuinely ready for a workflow, and what to do when it isn't yet. The assessment side looks at a small set of practical criteria: how often a task happens, how clearly its rules can be defined, how much room there is for error, and what its data flow and security profile looks like, where the data comes from, what it touches, and where a person needs to stay involved. The redesign side is where people most often get stuck, because they treat the first answer as final. In practice, a task that fails the fit test can often be restructured: separating the mechanical steps from the judgment calls, defining a clear rule for handling exceptions, or standardizing an input so a workflow has something consistent to work from. Oliviana will walk through one real example, applying both parts of this process in front of the room. This session is a preview of two parts of AI4ALL's full-day AI Workflow Builder workshop, October 6 at Equals in Amsterdam, part of World AI Week.