While machine learning can be integrated into a number of different processes, the path to integration is similar across the board. Initially using a corpus of known materials, the machine-driven processes record these pilot examples. From there the models use the data to provide suggestions to human processors who make the final determinations. The machines learn from these user selections and over time the processes reach sufficient confidence and accuracy to perform the process themselves, automating certain cases. The final step comes with the machine models taking on the bulk of the work in an automated fashion only allowing the rare edge cases and the routine audit level of documents to post for human review.