In World War II, the Allies successfully cracked secret German codes through arduous analysis. Today, financial institutions (FIs) have artificial intelligence (AI) to break a different kind of code: the money laundering code. Criminal organizations are growing in sophistication, talent, and persistence. These organizations often fund their operations through money laundering, and therefore need to frequently change the methods and the details of their money laundering schemes to thwart conventional anti-money laundering (AML) defenses. AI and machine learning are accessible tools designed to assist FIs identify, investigate, report, and prevent AML threats.
In November 2016, the New York Department of Financial Services (NYDFS) fined the New York branch of a global bank $215 million for AML violations. One of the main overt acts cited in court documents was that upon the hiring of a New York compliance leader, the leader discovered that a large percentage of the Society of Worldwide Interbank Financial Telecommunications (SWIFT) messages transiting the New York brank had unidentified numeric codes preceded by letters. After discovering the codes, the global bank discounted the finding and sidelined the New York compliance team’s efforts to investigate. This compliance breakdown allowed international transfers to transit the New York branch without any sanctions or AML screening because the codes were identifiers for foreign customers.
AI solutions interfaced with existing FI systems can detect unique and suspicious codes immediately and flag transactions for further investigation. Although AI could not have changed the deceitful global culture at the sanctioned bank, AI could have assisted New York compliance leaders to identify, isolate, and report the coded transactions immediately. AI also has the unique ability to analyze the numeric codes for similarities and clues that would take human investigators an exorbitant amount of time to complete.
Trade-based money laundering is an emerging threat for AML professionals. According to the World Trade Organization, international merchandise exports totaled approximately $16 trillion in 2015. Trade finance transactions with invoices and downstream payments to exporters, freight forwarders, customs brokers, and other merchants are staples of international trade. Trade-based money laundering organizations routinely falsify invoices to cover their exports. Often, the organizations sloppily create duplicate or non-consecutive invoice numbers, which is a key red flag for law enforcement investigators. Recent findings by QuantaVerse data scientists located several examples of shippers simply inverting numbers on inconsistent digits of invoices to give the appearance of legitimate trade transactions.
The Originator to Beneficiary Information (OBI) section of transfers hold a treasure trove of information for the AML investigator and law enforcement. Astonishingly, money launderers will list innocent clues about the transfers, their accomplices, and their intent in the OBI fields. Financial criminals also routinely use common addresses of co-conspirators interchangeably to deflect KYC and investigative efforts.
AI can instantly detect OBI clues through dynamic keyword searches and comparison of OBI messages with other transactional data. Additionally, AI analyzes addresses in the transfer data to identify previously used addresses linked to other customers or accounts. This enhanced data analytics solution is essential for FIs to maintain proactive, tactical, and strategic AML monitoring and investigation programs.
AI is the right tool to combat the hidden codes of money laundering. Cryptic SWIFT messages, erroneous invoice numbers, OBI messages, duplicate/linked addresses, and other money laundering clues are all buried in the terabytes of transactional data that FIs are responsible for monitoring. AI has the ability to effectively, and consistently analyze these and other data points for actionable intelligence. AI solves the problem and drives efficiency.