If we want to explore alternative solutions, I’d rather do that before implementing one of them.

Sounds cute. But you won't know what any of them really are until you've built at least one of them. You can write specs and have meetings for years and you'll still miss something fundamental about the approach that will be discovered in the first hour of building.