Are you looking to use the same scoring model with different thresholds, or entirely different scoring for activities? If you're not certain, I'd recommend isolating only the threshold or the model but not both, as it will be simpler to maintain.
If your scoring is global and your thresholds are local, typically you'd see specific campaigns where the thresholds are the trigger criteria, and you'll have one of these for each segment. If you add in the RCA complexity you may need separate models for each segment.