My take away is that the overall labor per customer decreases because workers are spending less time on non-cooking/serving tasks. The paper implies the efficiency gain is in how labor is used rather than production or serving efficiency. This seems like an anticlimactic conclusion.