Yes, these sorts of tasks are where LLM's are exceedingly useful.

But I was talking specifically about coding agents.

(A.k.a. spend four hours micromanaging prompts and contexts to do what can be done in 15 minutes manually.)