Because we would never ship
champions anything if we waited for things to be perfect. That's a larger take on the question, but and we do explore if we can do a champion before we start working on them. Famously Veigo was a champion we started on but ran into unforeseen constraints that forced us to put him in the ice box.
More seriously to your point though, we don't always know what is needed for a champion when we start working on them. Particularly the kinds of things that come from working on a champion and after having played them a bit. Things like Lillia and Zoe's third abilities (E) drawing a line on the mini-map were not things we thought of when we started on the champions, but that we discovered would be good to have through playtesting. Lillia's R showing the number of enemy champions that it would affect didn't show up until very late in her development.
Other times something ends up being easier than we thought, or even more difficult, maybe someone leaves the team or they need to shift their focus to another area of the game and we can't do everything we want to.
This doesn't mean we would lower our quality bar and release something we are not confident in, we do feel Zoe is in a good place. Would this be an improvement? Possibly, we haven't gotten to test it, but this is where iteration and experience help us in the future.