A new automated system has been proposed to thoroughly review tax returns of public and private companies making over a billion in revenue, generating an expected $67 billion in the first five years and an additional $20 billion in the following five. The system aims to prevent tax fraud and misrepresentation and utilize machine learning to identify exceptions.