The aim of the project was to develop a system for detecting fraud in the loss adjustment process in the sales network. A rule design and processing engine was designed and implemented to detect potential fraud before it is financially materialized. The processing results have been integrated with the notification platform, reporting system and business process.
Solution
Complexity:
20 TB
Dozens of AI algorithms Technologies used:
Microsoft SQL Server (database, SSIS, SSRS)
Drools
R
Result
Implementation of machine learning algorithms that significantly reduced the risk of abuse
Implementation of a platform that allows business users to manage the fraud detection system
System integration with data warehouse and domain systems