Yes - but Amphastar's counter is that that is a plot of a CE curve. How do you know what it corresponds to in the HPLC curve (which is what they use)? And is 'separation' the same as 'determining it is important thing'.
Neither HPLC or CE is a "destructive" method, and therefore once you load a sample onto the assay column, you can collect the peaks that are coming off at various times. Those collected peaks can then be examined by NMR or whatever other method you choose.
As you're developing a process for production, HPLC or CE analyses at a given step should provide the same trace from batch to batch as evidence of reproducibility. If, in the development process, you've already determined what each of the individual peaks are from a given HPLC / CE analysis (through NMR), then I think it reasonable to suggest that in subsequent batches that you can know definitively what each peak is without having to individually re-analyze them.