We argue that classical transitions can be the key to explaining the long standing puzzle of the fast A-B phase transition observed in superfluid Helium 3 while standard theory expects it to be unobservably slow. Collisions between domain walls are shown to be capable of reaching phases inaccessible through homogenous nucleation on the measured timescales. We demonstrate qualitative agreements with prior observations and provide a definite, distinctive prediction that could be verified through future experiments or, perhaps, a specific analysis of existing data.