Has anyone changed the way they use Wan 2.7?

I used to treat Wan 2.7 AI Video Generator like a one-click video tool.

Lately, I've started using it more as an iterative workflow—generate a rough version first, review the motion and pacing, then refine only the parts that need improvement.

It's made the whole process feel much more predictable than trying to get the perfect result in one generation. Curious if anyone else has shifted to this approach.