Oklahoma Employment Security Commission: Reducing unemployment claims fraud
The Oklahoma Employment Security Commission is an independent agency responsible for providing employment services to the state’s residents. The commission is part of a national network of employment service agencies and is funded by the U.S. Department of Labor.
The Covid-19 pandemic set off a surge of unemployment claims, and the Oklahoma Employment Security Commission (OESC) needed a better way to review these claims for possible errors and fraud. It turned to Egen for help to build an intelligent solution.
Alongside the uptick in claims volume, the OESC was grappling with implementing new requirements and trying to update inefficient legacy systems. It created a perfect storm to increase errors and fraud, putting more burden on the OESC’s human adjudicators and meaning benefits might not reach those who truly needed them in a timely manner.
To reduce the risk of improper payments and efficiently meet the requirements, state agencies like the OESC needed an automated way to analyze claims and identify potential issues. Egen created a solution bringing together machine learning, artificial intelligence, and Google Cloud tools to uncover suspicious patterns, detect fraud, and improve verification.
OESC wanted to develop a solution as quickly as possible to better serve people in need. The Egen team started by sitting down with OESC adjudicators to discuss the most important data sources for finding clues about suspicious activity. Egen then combined multiple data streams from within OESC and external sources to generate risk scoring that flags abnormalities and possibly fraudulent claims for review by human adjudicators. The solution also ingests data in visualization tools such as Looker and Data Studio so adjudicators can quickly analyze information and spot trends.
Knowing how important it was to allow OESC to focus on processing real claims, Egen had a rapid development environment and worked closely with OESC team members to bring in their insights every step of the way. With an automated solution to detect anomalies and generate risk scores for claims fraud, human adjudicators can focus on higher priority cases that have a higher chance of being fraudulent.
Now, fraud and improper payments are found more efficiently and effectively, and the people who really need the benefits can receive them faster.