Accurate and Fast Anomaly Detection for Anti-Money Laundering

Mitigate your organization’s risk and ensure compliance with augmented anomaly detection

For banks and other financial services who must comply with increasingly stringent regulations across the globe, money laundering, and detecting it, threatens to become a crisis — one which already costs the global economy nearly $3 trillion a year. To combat this, companies need better data and new methods to spot anomalous transactions that could aid anti-money laundering (AML).
The problem

Our customer tracks millions of financial transactions every day and must comply with local and international AML regulations that require them to report any abnormal behavior. A rules-based system was in place but they still relied on looking only for the same tried-and-true indicators that may no longer be relevant due to changing tactics and more sophisticated money laundering methods.

The solution

Our customer focused on building indicators using external data from the Explorium Enrichment Catalog that could point to troublesome transactions even on seemingly unrelated data. One of the big challenges they faced was meeting regulations about the use of personally identifiable information (PII). Our customer needed to find data that could highlight potentially illegal transactions without using data that could run them afoul of international data privacy regulations. To do this, they use Explorium to connect to a variety of contextual data sources to build new features, such as: 

  • Origin data combined with personal information to determine whether a client is a real person
  • Demographic data by region compared to economic information to point to abnormalities
  • Crime statistic data connected with individual transactions based on point of origin
  • Lending information and financial stability data connected to historical risk scores
  • Financial risk scores based on the volume of transactions and transaction amounts
The results:

Faster anomaly detection for lowered compliance risk

The new anomaly detection model helped our customer better understand and identify anomalous transactions. While not all anomalies point to money laundering, the more precise detection tools allowed them to cut down on the time they spend identifying and examining transactions that are flagged. Additionally, the company was able to improve their accuracy in detecting money laundering transactions by 15%, to 89% overall. The result was savings in their compliance efforts, both in money and time spent on each transaction. More importantly,  they were able to offer faster services to customers without sacrificing security.

Schedule a call and learn how to fight money laundering with enriched anomaly detection models.

Other resources you might be interested in

We're Hiring! Join our global family of passionate and talented professionals as we define the future of data science. Learn More