Automated process execution based on evaluation of machine learning models

This invention describes a computer system that learns how to automate business tasks. It does this by looking at past data to build a machine learning model, specifically using decision trees, which are like a series of 'if-then' rules. The system then pulls out these rules, checks how accurate and comprehensive they are, and uses the best ones to automatically run parts of a business process when certain conditions are met.