The Arizona Department of Economic Security (AZ DES) works with residents to ensure that every child, adult and family in Arizona is safe and economically secure. The AZ DES is responsible for managing the Pandemic Unemployment Assistance program for the state.
Following the passage of the Coronavirus Aid Relief and Economic Security (CARES) Act in 2020, agencies across the country saw a huge surge in unemployment claims, specifically the newly created Pandemic Unemployment Assistance (PUA) benefit. With millions of requests for assistance coming through its PUA portal, the DES was under mounting pressure to release payments from its backlog of UI claimants, but it suspected that many of the claims being submitted were fraudulent.
Using the Google Cloud Platform, SpringML was able to develop a framework to segment potentially fraudulent claims from legitimate ones.
Using analytical techniques we built a list of flags to help determine fraud and developed a rules-based system that generated a score for a particular claimant. The flags were weighted based on their importance in detecting suspicious activity. SpringML coordinated with DES AZ personnel that conducted audits and optimized the flags and weightings based on feedback provided.
This enabled the DES to release payments for claims with a low probability of fraud, identify those with high indicators of fraud, and focus their adjudicators on investigating the more complex claims.
Our solution provides a verification process to facilitate automation and reduce the amount of time to check claims. The scalable system was able to process 4.3 million claimants in a matter of hours.