What are some of your proudest achievements from the past year? We are particularly proud with the increased performance of our models and our ability to show that compliance practices can be constructed in a way to inform models. For example, models to predict which third-party vendors interface with governments have improved from 4-5% accuracy to over 80% accuracy in many jurisdictions. Models to predict graft and evasion of controls also likely experienced over 20-times improvement in the three years that the program has been running. The project has dramatically reduced investigation costs and provided insights into compliance risks that allow for the effective prioritization of resources.