As the title says, allowing audio inputs on redux would allow reusing Redux Fx, doofers, etc. It could be implemented as a sample on the waveform or as a whole instance thing.
Extra points if patterns (phrases), that ignore notes and commands, were available for envelopes, pan and filtering. Then it could come pretty near replacing CamelSpace.
PS: I cram this here, but what stops me using Redux as my only sampler is that velocity can’t modulate the sample start point. This is a 101 feature on a sampler IMO. It limits its usage as drum sampler.