Case study: Fraud Monitoring and Prevention for a Leading US Bank

About the Client

The client is a leading US based, diversified financial services company and provides retail, commercial, and corporate banking services to individuals, businesses, and institutions.

Business Context

The client’s objective was to detect card fraud instances like skimming, “card not present” and identity theft. As fraud patterns were shifting sharply, with greater proportion of transactions being done online, their existing solution had higher false positives in certain segments and lower capture rate, leading to increased fraud losses.

BRIDGEi2i Solution

BRIDGEi2i partnered with the client to refine & implement a new fraud detection solution with an aim to reduce fraud losses without impacting customer experience.

BRIDGEi2i leveraged its data engineering expertise to used pre-designed ETLs and API to read, extract, transform and load data. Data used ranged from customer demographics, purchase patterns of the customer, bureau information, distance features from geo-location codes to socio-macro- economic indicators

BRIDGEi2i’s AI labs built an un-supervised model for personalized anomaly detection. The ensemble of various models was carried out to minimize false positive, through a dynamically defined cut off.

BRIDGEi2i leveraged its proprietary AI Accelerator ‘Watch Tower‘ to implement an Early Warning Fraud Alert system for real time detection of suspicious and Fraudulent transactions as well as recommendation of mitigative actions.

Business Impact

Fraud Detection Rate
False Positive Rate