I disagree. "Premises first, then conclusions" is a useful abstraction, but here's another one.
Logic is spacial. Premises at the bottom, with conclusions built above them. Conclusions are "based on" premises; conclusions must "rest on" sound premises.
You can, in fact, examine the conclusions first, but they will only be as sound as the underlying premises.
Don't make the mistake of believing that because we usually state premises first, that there is any sort of mandatory temporal relationship between premises and conclusions. It's a useful convention, it's convenient for verifying soundness, but logic certainly doesn't require it.