For RI Regulatory Labelling, connecting segments at alignment as either as n:1 or 2:2 causes issues further in the process with PerfectMatch, when due to source edits such connections un-merge and more leverage than necessary is lost. There is a check that catches n:1 connections at Alignment, which helps a lot, but it would be a good enhancement to have a special Regulatory Labelling option whereby no n:1 or 2:2 connections were suggested by the automatic pre-alignment in the first place, to save linguists time they spend checking and reconnecting. Possibility to align n:1 or 2:2 manually should still remain for some exception. As for the option to split source segments - it should also still remain as there might be cases where it is absolutely necessary, but there should be an automated check to find any split segments (maybe similarly to what's available in Studio Advanced Filter).