What often happens next is the realization that the existing system was handling far more edge cases than it initially appears. You often discover these edge cases when the new system is deployed and someone complains about their use case breaking.

The reverse is also sometimes true and it's when a rewrite is justifyable.

I've worked with many systems that piled up a ton of edge cases handling for things that are no longer possible, it makes the code way harder to follow than it should.

I've had successful rewrites that used 10x+ less the amount of code, for more features and significantly more reliable. And completely eliminated many of the edge cases by design.