There should be a way to re-assign the speaker labels from scratch; any time I make a mistake identifying who's speaking, the deed is done and the entire transcript is locked into the wrong speaker configuration. Just being able to rename all the incorrect labels in one go - that is needed.