Background
Client is one of America's largest financial technology and bank holding company headquartered in Austin. It is the world's largest prepaid debit card company by market capitalization. They also act as payments platform company and is the technology platform used by Apple Pay Cash, Uber, and Intuit.
Challenge
Client faced several challenges and financial risks resulting from obscured transaction datasets, making the actual nature of the data hard to determine. Their obscured datasets led to false positives that affected actual customer transactional data and customer experience. The Bank also faced challenges with predictive analytics, as it created dummy datasets to match real-life scenarios concerning the highest suspicious activity observed by the bank.
Solution
Our team built an anti-money laundering system using a machine learning (ML) Challenger Model, using unsupervised learning in parallel with Actimize. Our team also developed Dashboards linked to generated alerts, related to Suspicious Activity Report (SAR) generation.
Outcome
Our implementation with the Bank had several positive outcomes. We created a product for them to detect fraudulent transactions. The use of this tool decreased transaction response and prevention times. Client also saw a ~20% reduction in false positives leading to customer satisfaction alongside improvement in data fidelity using the ML fuzzy logic technique.
Background
The Green Dot Corporation is an American financial technology and bank holding company headquartered in Austin. It is the world's largest prepaid debit card company by market capitalization. Green Dot is also a payments platform company and is the technology platform used by Apple Pay Cash, Uber, and Intuit.
Challenge
There were several challenges and financial risks resulting from Transaction data sets that were obscured, so the actual nature of data was hard to determine. This led to false positives affecting actual customer transactional data and customer experience. Also, predictive analytics was challenging as creating a dummy dataset to match real life scenarios with respect to the highest suspicious activity observed by the bank was difficult.
Niha’s Solution
Niha's team was hired to provide services for development and implementation of an anti money laundering system. We developed the ML Challenger Model using unsupervised learning in parallel to Actimize. Our team also developed Dashboards linked to the generated alerts which were further linked to Suspicious Activity Report(SAR) generation.
Outcome
Greendot Bank had several positive outcomes, including:
- Detection of fraudulent transaction detection using a product our team developed, thus decreasing response and prevention times.
- ~20% reduction in false positives leading to overall improved customer satisfaction.
- ML fuzzy logic helped with improving the data fidelity.