PwC assisted a leading payments company with conducting an investigation to identify and examine suspicious customers in accordance with anti-money laundering regulations. Large volumes of customer and transaction data and no prior analytical expertise in this field created significant challenges for the client.
Enhanced traditional AML rules by uncovering instances of suspicious account activities and behaviors hidden within terabytes of transaction and customer data.
Developed graph search tool to identify all direct and indirect customer and product connections for each customer - explored 90M entities and 180M relationships to quantify customer account structures and identify risky topologies.
Analyzed a data set of close to 1T transactions to develop a metric to identify suspicious customer and merchant relationships based on relationship strength relative to other cardholders.
Matched customer transaction sequences to AML risk patterns that capture specific instances of risky behavior, eliminating false positives.
- Developed an algorithm to match duplicate customer entries caused by name and address changes as well as duplicate account applications - eliminating over 10K duplicate customers records.
Impact on client’s business
Reduced the number of false positives to a quarter of what was initially estimated by applying advanced analytical techniques such as nPath, network, and segmentation.
- Reduced the time required for manual investigations through increased visibility into the suspicious activities and customer/account relationships.