Defeating the Dynamic Money Laundering Culture with Data Science

We Have a Problem

It is common knowledge how much “dirty” money is laundered every year. We have all read the United Nations Office on Drugs and Crime’s (UNODC) 2009 report, which estimated that dirty money (criminal proceeds) accounted for approximately 3.6% of the global GDP and estimated that about 2.7% ($1.6 trillion) was laundered through various methods. Drug cartels, human traffickers, organized crime, wildlife traffickers, and other criminal groups are now commonly referred to as transnational organized criminal organizations (TCO) and are now making successful money laundering operations integral parts of their business plans.

Financial institutions rely on transaction monitoring systems (TMS) to screen transactions for suspicious activity using static rules that require constant tuning and testing. Financial institutions are struggling to handle 95-99% false positive rates that arise from the varied rules they are using, and subsequently allocate considerable employee resources to clear the false positives and investigate the real/serious alerts.

Working Smarter to Defeat Money Launders and TCO

How can we all work together in a smarter and more efficient way to defeat money launderers? Communication, knowledge sharing, technology, and professional passion are key. When it comes to technology advancements, AI is one solution that works well and should be embraced. QuantaVerse is helping clients proactively defeat criminals through a number of cornerstone solutions working in four areas:

• Pre-TMS Entity Resolution & Risk Scoring reduces transaction monitoring system (TMS) false positive alerts by classifying the risk of each transacting party, down to the pseudo-client level.
• QuantaVerse Alert Investigator automates investigations when an alert is produced by a TMS.
• QuantaVerse False Negative Identification and Investigator examines transactions for financial crime risks that have been missed by a TMS.
• QuantaVerse Third Party Risk Management provides risk managers with better insights into the variances in account activity that might indicate risks of financial crimes, or that suggest an account is being used for something other than its stated purpose.

AI as a Detection Agent Complementing TMS

Since its inception in 2014 and after analyzing petabytes of data, QuantaVerse’s AI solutions have detected scores of financial crimes in the areas of human trafficking, international financial fraud, terrorism financing, drug trafficking and all types of money laundering (trade-based, cryptocurrency, transactional, trade finance).

While supporting and enhancing clients’ TMS solutions, transactional data was analyzed using AI. The analysis was able to detect different anomalies that the rules-based engines were not properly tuned to detect. In each typology, standard money laundering AI agents that somewhat mirror TMS rules were employed, but in each scenario, additional proprietary and unique AI agents were utilized to provide missing information that the client TMS solutions were not able to execute.

• Human Trafficking
  • Linked KYC data used to detect customer/criminal commonalities
  • Velocity and timing of transactions
  • Missing normal peer customer transaction types, ie. salary payments
  • Media analysis

• International Financial Fraud
  • Analysis of message/OBI & BBI fields
  • Invoice & transaction amount analysis
  • Adverse media
  • Address analysis & validation
  • Directional payment analysis
  • Geographic targeted analysis for high-risk fraud jurisdictions

• Cryptocurrency Money Laundering
  • Analysis of message/OBI & BBI fields
  • Media analysis
  • Address research, analysis, & validation

• Trade-Based Money Laundering
  • Analysis of message/OBI & BBI fields
  • Invoice & transaction amount analysis
  • Media analysis
  • Directional payment analysis
  • Address analysis & validation

• Trade Finance Fraud & Money Laundering
  • Analysis of BOL (Bill of Lading), invoice, and related trade document information for validity
  • Analysis of message/OBI & BBI fields
  • Invoice & transaction amount analysis
  • Directional payment analysis
  • Adverse media
  • Address analysis & validation

• Terrorism Financing
  • Analysis of message/OBI & BBI fields
  • Proprietary and dynamic keyword analysis
  • Media analysis
  • Geographic focused analysis
  • Charity, crowdfunding, and product purchase analysis

• Drug Trafficking
  • HIDTA/HIFCA and other geographic analysis
  • Analysis of message/OBI & BBI fields
  • Proprietary and dynamic keyword analysis
  • Directional payment analysis
  • Adverse media

• Transaction Money Laundering
  • Identification of Web-based vendors/merchants
  • Various money laundering transaction typology analysis
  • Web research
  • Identification of irregular payment activity

• Shell Company UBO (Ultimate Beneficial Ownership)
  • Advanced web research
  • AI agents create a link analysis of relationships
  • Third-party risk identified and graphically depicted

One example of AI’s ability to determine economic purpose and risk rate was exemplified when QuantaVerse reviewed approximately 2.2 million transactions for a global financial institution. The QuantaVerse AI Financial Crime Platform detected payments between a casino in Asia and computer manufacturer in Europe. This relationship on the surface made sense. There is nothing unusual about a casino doing business with a computer manufacturer. However, when the platform examined the payment flow, it discovered that the computer business was sending large transactions to the casino, which did not make sense. Further, the AI platform revealed that the computer manufacturer’s owner was involved in a tax fraud case and was reportedly liquidating assets offshore. The case was flagged for further investigation by the client.

Whether you are working in a bank or in law enforcement, you know that money laundering has drastically changed over the years. From the Cocaine Cowboys era during the 1970-1980s in South Florida where trading bags of cash for cocaine and million-dollar cash real estate deals were common to ISIS supporters selling bogus goods on Internet sites, one common element remains: crooks are smart and learn from each other’s successes and mistakes. Crooks learn each time a financial institution closes their accounts due to suspicious activity, akin to the military conducting an after-action debriefing after every operation. While there is no silver bullet for preventing, detecting, and reporting all financial crime, advanced AI and data science give us an edge to stay competitive with the modern, evolving criminal element.