Thinking about it doing a conversation would be simplicity itself using the editable timeline. simply load the first figure into mimic with the full conversation and delete the phonemes of the second figure from the timeline. saving the pose then do the same with the second figure, deleteing the phonems of the first figure.