we end up with either multiple maxims to test (and the possibility of conflicting results) or a single “relevant” maxim... [universalisability] should not be viewed as testing the actual maxims on which agents might act in some specific context... [we] should be seen as working with “generic” maxims, which can apply to many different actions, and more than one of which may apply in a specific instance (Herman 1996, ch. 7).
So Kant does not prove "maxims on which agents might act... [because?] more than one... may apply in a specific instance"
So how do we obey the moral law? I may have misunderstood, in which case, given you can understand how, can you explain why?