Graph Networks

Case Study

Business Problem: This large financial services customer was implementing a compliance and AML solution with many downstream systems in the ecosystem integrating with various vended and in-house built systems. The challenge in one of the investigative systems was not being able to quickly turn around processing high severity alerts from the monitoring systems. The reason for the high turnaround times varied from integration issues to drilling down to other systems for customer/account relationships. Because of the latency in resolving alerts on time, there was significant impact to business relationships.

Current State of Analytics Maturity: The customer at the time of engagement were highly matured in their analytics systems both in terms of reporting and data access. There were various tools, technologies and processes established and in production. However, in the area of visualization, advanced analytics and machine learning there was room for improvements in terms of elevating some systems from POC to production and realizing ROI.

Solution:

Re-designed customer data in an RDBMS setting to a Graph Network Model to receive alerts on fraud and money laundering
Developed dashboards and graphical workflows on top of the graph networks to enable analysts to view alerts propagated by upstream systems
Business Benefits: Helped build a new revenue stream and cost savings over vended product usage
Business Challenges: Current vended products providing partial and inflexible solution. Custom build out was a hard but successful sell

Outcomes:

Over 40% cost savings and incorporated best practices for agile analytics development
Fully trained business teams (with our proprietary analytics framework) in terms of executing Analytics projects and managing quasi technical teams