False Positive Reduction for Transaction Screening and Customer OnboardingBFSI
Transaction screening and customer onboarding are important parts of the compliance process. Having a poor screening and onboarding process could lead to high risk and penalties from regulatory bodies. A false positive is a critical performance indicator of a model that helps in balancing the risk scale.
Having a higher false positive will lead to additional investigation, increased cost, and affect the customer experience with unnecessarily blocked transactions and a slow onboarding process impacting the business. On the contrary, having a low false positive may lead to onboarding high-risk or potentially unnoticed money laundering activities.
The client, a leading remittance house in the GCC region, sought a real-time names and transaction screening solution to mitigate risk and provide a safer financial space for their customers.
Challenges:
- The client lacked a benchmark for the accuracy of name-matching algorithms.
- Fine-tuning the model to reduce false positives and screening time without compromising on recall.
- Developed a machine learning driven module that improves the name matching algorithm by optimizing the false positive while retaining the recall rate.
- Implemented a real-time solution that manages all the algorithms and generates the response for a payment screening within less than 1,000 milliseconds.
Delivered Fine tuned model helped client to reduce false positives and investigation time. Reducing the time to mitigate risk better and streamline their name and transaction screening.
Key Benefits:
- Market-leading Accuracy with 24% less false positive
- Enhanced Transaction Security for Customers.
-
Streamlined Customer Onboarding
-
Improved Operational Efficiency