I believe the final destination of every machine learning system, every recommendation engine, and every data science initiative should be the auto-play.