The front line of the fight against money laundering and financial crime includes AML investigators and analysts, government investigators, regulators, and the companies that are developing innovative technologies to combat financial crime. Together, this collaborative team is better able to stop money laundering by implementing solutions such as artificial intelligence (AI) and machine learning for enhanced monitoring, investigation and KYC.
Traditional transaction monitoring (TM) systems are a necessity in the financial industry. Financial institutions (FIs) rely on TM systems to target high-risk indicators of AML typologies such as:
- Funnel account activity: When an account receives a high volume of deposits or transfer activity and then rapidly transfers the funds to another account, often in another geographic area. These accounts are common with human trafficking organizations.
- High velocity activity: When an account exhibits an abnormal amount of activity over a short timeframe that is not consistent with similar customers or accounts. These accounts are common with various money laundering schemes for all types of criminal organizations.
- Routing of transfers through multiple jurisdictions: Entities create business relationships or subsidiary companies in many countries allowing them to route money through multiple jurisdictions. Some account owners often take a small deduction off the top (10-20%) to make it difficult for transactions to be tracked across relationships and locations. Running transfers through secrecy havens further complicates tracking of illicit funds.
Other money laundering indicators that TM systems regularly monitor include:
- activity outside of the customer’s expected account profile;
- structured transactions;
- round-dollar transactions;
- pass through accounts;
- many to one transaction flows;
- one to many transaction flows;
- foreign fighter typology red flags; and
- tax amnesty & tax avoidance cues.
Flagging questionable transactions is invaluable, but TM systems depend on rule-based scenarios that require continuous tuning. This is costly, time-consuming and allows bad actors to possibly execute money laundering schemes for months before the system has been tuned to identify their latest behavior. A properly implemented AI-based financial crime solution can supplement a TM system creating an integrated financial crimes risk prevention, detection, investigation and mitigation solution.
For the overworked compliance team, an AI-based solution can collate massive amounts of customer information from many sources: TM systems, KYC databases, Lines of Business (LOB) customer information, the deep web as well as open source investigative and Internet sources. Once collated, the AI-based solution can send out AI agents to query all data points for red flag indicators and report back to a machine learning component of the system. The machine learning box can then begin to constantly query enterprise FI data to learn where the customers’ good and bad behavior exists and report the anomalous behavior as highly qualified alerts to the investigative team. Equally important is AI’s ability to monitor customers’ relationships to other customers and entities and learn from their associated behavior.
One of the costliest and time-consuming yet vital part of any FI’s compliance program is the KYC system. A properly installed AI-enhanced machine learning solution supplements a KYC program by providing constant and deep monitoring of customers’ transactional and financial profile through FI systems and other investigative resources. The end result is a cost-effective enhancement to the KYC process which provides a 360-degree customer profile.
Consider how round dollar transactions would be handled in an AI-enabled environment. FIs are warned that these types of transactions are telltales of illicit drug payments, however, legitimate “round dollar” transactions are also very common. An AI-enhanced system can instantaneously check the round-dollar transaction flags from the TM system against other indicators such as non-complementary lines of business, KYC data, history of transactions between entities and structuring behaviors to accurately eliminate false positives. This will not only reduce the burden on investigative teams, but it will also document more robust data to be analyzed.