While human scoring of agent conversations has been the standard for decades, it is no longer necessary for humans to pour over countless hours of recordings to unearth valuable insights.