In the past few years, we have seen increasing pressure on compliance officers to investigate for new money laundering schemes. Many banks addressed this by putting more effort into working alerts.
This is not the best approach to AML case management, as this example illustrates: The Swiss Financial Market Supervisory Authority (FINMA) internally maintains a “Red List” listing some 20 Swiss banks are listed that the supervisors believe are most likely to have customers who might be involved in crooked business — specifically, money laundering. Recently, a Tier1 bank in Switzerland hired 800 investigators to work on alerts — this appears not to have substantially lowered the risk, and it’s likely that FINMA might add the bank to the list mentioned above. That could cause dramatic damage to the reputation of the bank.
Clearly, we need other approaches to AML case management than hiring investigators. An increase in both effectiveness and efficiency of the overall compliance process is required. We believe the best way to combat this problem is:
- Using artificial intelligence and machine learning for a prioritized alert generation process
- Using robotics to effectively dispatch alerts to the investigation team and carry out auto-decisioning wherever possible
An Industry-Wide Approach
“During the early 1990s, the Toyota Production System (TPS) became widely accepted across the industrialized world. Companies from different industries gained in experience by applying lean principles as well as by developing and refining customized approaches to optimize products and processes.” (http://sciencealerts.org/aiem/wp-content/uploads/downloads/2013/04/6-J1-10320-Final-Paper-AIEM.pdf)
The Toyota Production System for shop floor management helped the manufacturing industry increase productivity and quality, while optimizing organizational structures, such as centers of competence. This was an industry-wide approach to automation that also improved the work done by individuals.
This is the approach we need to be taking to AML case management, to improve results while reducing the amount of time and investigators needed to review high volumes of false positives.
Such an approach would involve multiple steps:
- Develop rules and use AI to automatically tune the result set. Repeated matches and scenario hits which have been classified as “false positives” should be avoided. Also, similar patterns compared to previously decided “false positives” should be auto-dropped if possible and allowed by the regulator.
- Set up a dedicated multi-step organizational structure to dispatch alerts to the right skill level. An increased automatization is a must-have:
- An initial analysis can be done by standard (low-skilled) staff. Standard situations can be resolved either by robotics or by standard procedures and clearly defined tasks (executed by the low-skilled staff).
- Cases requiring further investigation need to be forwarded to money laundering subject matter experts. They will apply their expertise and provide a view on typical money laundering schemes.
- Provide the workflow capability to that ensures team cooperation. The solution must follow the structure of the compliance organization, not vice versa.
- Control and monitor the process. Dashboards displaying the right KPIs will drive efficiency and provide the full picture of the risk situation of the financial institution and the status of each alert.
The Regulator’s View
At meetings with bank regulators this month, we presented Siron®ACM, FICO TONBELLER’s component for central alerts and AML case management. Representatives of the regulators emphasized that from their point of view, it’s critical for financial institutions to:
- Work on alerts within a short period to make a decision and, in case of a confirmed suspicion, to file a suspicious activity report for the FIU.
- Have all processes documented in an audit-secure way, so the auditor can reproduce each step of the decision process.
- Consider the 360-degree view on alerts.
This is the approach we have taken with our Siron®Anti-Financial Crimes Solutions. Our alert and case management Siron®ACM provides a 360-degree view on a customer (from various alert generating systems), robotics and a configurable workflow, as well as a plug-in for artificial intelligence and machine leaning that leverages FICO’s patented innovations in anomaly detection, which have been used in the fraud space for decades.
For further details please refer to http://www.tonbeller.com/en/solutions/anti-money-laundering/anti-money-laundering-for-banks/.