I’ve been looking at this AI copilot idea where you basically describe what you want in plain English and it generates the workflow for you. On the surface it sounds amazing—no coding, instant turnaround. But I’m trying to understand the realistic workflow here.
When something like that generates code or a workflow structure from natural language, it’s not going to be perfect the first time, right? There’s got to be debugging, adjustments, edge cases you didn’t think to mention in your description. I’m trying to budget for actual project time, and I need to know: are we talking 5% rework, 30% rework, or something in between?
Has anyone actually used an AI copilot to generate workflows from plain English descriptions? What percentage of the generated output do you actually deploy as-is versus having to rebuild or significantly tweak? I’m trying to calculate whether this actually saves development time or just shifts where the effort goes.