◇ Could this help me?
Yes, and the method generalizes well beyond design. Mapping a workflow before automating any of it is the step almost everyone skips, and it is what prevents the common outcome of automating whatever was easiest to automate rather than what was actually costing time. The two-resolution idea is the clever part: a high-level map is what makes the change reusable across projects, while a specific project map gives you something concrete to test it against. General practice, with one honest caution about the framing that exploration is waste, since some apparently unproductive iteration is how a hard problem gets understood, and the distinction worth drawing is between exploration that teaches you something and repetition that does not.
Help me decide where AI actually belongs in my workflow instead of adopting tools at random. First, interview me and map how I work at two levels: a general version I could apply across projects, and the specific version of one recent project. Then help me score each step objectively on two axes, how much time it consumes and how much value it adds to what I deliver, and be willing to push back when I am defending a step out of habit rather than results. Give me the ranked list of candidates for automation, and for the top one, propose how to test the change against that specific project before I adopt it generally.