I wouldn't consider it an unreasonable goal when it's been the standard nearly all previous pre-ai re projects have held themselves to.