Reducing Fraud

•WEF has estimated 49% of global organisations have been a victim of fraud or economic crime.

•Forbes has estimated some banks spend up to US$500 million PA to improve and manage their Know-Your-Customer (KYC) and Anti- Money Laundering (AML) processes.

•KPMG estimates the average bank spends around US$48 million per year. In the US alone, total bank spending is more US$25 billion+ a year on AML.

Objectives

A major world bank created a new Financial Crime threat mitigation team.  Their mission was to proactively identify money laundering and other financial crimes such as sanction violations, illicit money use (weapons, human trafficking etc.).  Event detection in this area was compounded by the sheer scale of data (~40% of world trade is processed by this bank), the diversity of systems and countries (60+ countries, 200+ source systems, many from acquisitions) and ever changing nature of the frauds.

Solution

We conducted a 12 week data strategy assessment covering: culture, people, process, analytics capability, infrastructure. The findings of the study highlighted significant challenges in many aspects of how data was handled and specifically gaps in how machine learning approaches could be better used to detect fraud on a global scale.

Impact

●Exploit data assets: Ability to fully exploit business benefits of the Financial Intelligence Platform (FIPL) data repository.

●“Moving at the speed of crime” - created an innovative model deployment framework to facilitate faster and more agile deployment of new methods. Reducing time from months to days. 

© 2020 PS AI Labs