The use of artificial intelligence and machine learning, particularly neural networks, has effectively directed an increase in data-driven solutions and applications in different domains. Unfortunately, enterprise resource planning (ERP) vendors' emphasis on the importance of data integrity means ERP systems are less likely to protect themselves proactively through data analysis or by deploying external software with access to the data they maintain or process. Wholesale export of employee and business data into a data platform makes it possible for the ERP system and other business systems to use encryption and AI/ML software to proactively manage cybersecurity. This paper develops a conceptual framework for big data-driven ERP security management and uses it in six case studies. We found that applying a good enough simple learning model with the simplest network architecture will succeed in unmasking patterns that are more likely to have been overlooked. The case studies suggest that implementing big data-driven ERP cybersecurity will not significantly add software or maintenance costs, but harnessing artificial intelligence and machine learning services such as neural networks will add significant value to ERP customers.