Automated Invoice Control for Fraud Detection – DAS

Challenge

Nature of collaboration

To respond to this challenge, DAS partnered with Sagacify to develop an automated solution capable of identifying invoice amounts that appear abnormally high for the task and determining whether a dispute is necessary. The collaboration focused on aligning the system with DAS’ internal processes by integrating relevant claim data and ensuring that the model’s predictions reflected real-world billing practices.

What we built

Sagacify developed an AI-powered system capable of identifying invoices with abnormally high amounts for the corresponding task and determining whether a dispute is necessary. The model collects information from various internal databases, including details about the beneficiary, the parties involved, the invoice content, the billing history of the expert, and similar cases across different providers. This structured data enabled the model to make high-precision predictions. Given the use of highly sensitive data, understanding DAS’ business processes was essential to identify the most relevant features and ensure the solution integrated seamlessly into existing workflows.

Impact