The implementation of an automated fraud assessment script for an instant line of credit
firm resulted in 100% coverage of cases daily, encompassing all users, merchants, and
agents, while significantly reducing manual efforts to just a few hours per day.

The problem

The instant line of credit firm faced challenges in minimizing fraud by shop merchants and field agents onboarded on their app. Manual monitoring of transactions and activities led to verifying only 30% of fraud cases daily, increasing the risk of financial loss.

Methodology

An automated solution was implemented, extracting data from Amazon RDS and performing ETL using Glue. Data was modeled with Amazon Sagemaker for advanced analytics. Daily updates and visualization were done on Quicksight and also configured for automated reports which were sent to respective stakeholders.

Our impact

The automated fraud assessment script achieved near fully automated monitoring across users, merchants, and field agents, covering all cases daily. It significantly reduced manual efforts to just a few hours per day. This led to 100% daily coverage compared to the previous 30%, effectively minimizing the risk of fraud.