By just doing something manually 10-100 times, and collecting feedback, both understanding of the problem, possible solutions/specifications can evolve orders of magnitude better.
I think uncle bob or martin fowler said "don't buy a JIRA until you've done it with post-its for 3 months and you know exactly what workflow is best for your team"
Coding with English (prompting) is often most useful where existing ways of coding (an excel formula) can’t touch.
Using llms to evaluate things like an excel formulas instead of using excel doesn’t feel in the spirit of using this ai’s power.