Blog by Garima Chaudhary, Oracle Financial Services Financial Crime and Compliance Management Specialist.
On the timeline of financial services development, the anti-money laundering discipline does not have a long history. It started around 1985 and evolved in more recent years beyond an anti-crime, anti-drug orientation to include an anti-terrorist financing element post 9/11.
To have an effective anti-money laundering program, global banks and international bodies must deal with the confluence of data that emanates from the financial, social, and economic aspects of life. With recent technological advancements, tax evaders, terrorists, and cyber criminals hide behind and leave their trail in encrypted messages, social media, the cloud, big data, and the "Internet of Things." The focus for the past few years has been around unification of a financial crime and compliance platform, which brings various components of an anti-money laundering program as part of a single, unified compliance platform. Components include transaction monitoring, case management, suspicious transaction reporting, analytics, etc. as part of single platform. Several financial institutions have already jumped in and are now moving towards the next step of this journey. In addition to unification of a compliance platform, another major focus is on data and making sure there is quality data for detection and investigation for an effective program.
While the industry continues to work towards these aspects, there will be a next level of items that financial institutions will focus on over next few years:
- Machine Learning in Suspicious Pattern Detection: Traditional ways of query-based detection have been successful for suspicious transaction monitoring. This query-based detection allows financial institutions to adjust various monitoring parameters for much better accuracy. Query-based detection monitoring requires regular testing and updating, which in addition requires a massive quantity of human resources, technology, and money. Furthermore, query-based detection lacks the ability to apply better detection logic based on prior behaviors by itself. There is always a possibility to miss some suspicious behaviors due to delay in updates in the traditional detection.
And that is why this traditional query-based detection will evolve towards smarter systems, which will have any ability to learn by itself and keep up without human intervention. At a high level, learning will include historical customer behaviors and analyst conclusion to apply the updated detection logic. If it is possible to identify such repeated analyst conclusions and factor those in suspicious transaction monitoring, then the amount of false positives can also be dramatically reduced. This does not eliminate the need for regular verification of detection configuration; however, will certainly provide a smarter suspicious transaction detection capability.
- Robotics for Investigation: Cost of compliance in anti-money laundering due to huge investigation teams is increasing. There have been some recent fines not due to the inability to detect suspicious transactions, but due to the lack of investigation. Banamex USA acknowledged that they conducted fewer than 10 investigations and filed only 9 so-called suspicious activity reports — even though its monitoring system identified more than 18,000 transactions as “potentially suspicious” during that period (detailed article here).
Robotics is a general term that can refer to a few different uses of digital robots to automate work. In most cases, Robotics is referred to Robotic Process Automation (RPA), which is using Robotics to automate an entire process from start to finish. The other type of Robotics, which is discussed less often, is Robotic Desktop Automation (RDA), which combines human and robot. This technology allows an organization to automate many stages of an investigation process, allowing for more efficiency, consistency, and effectiveness. Financial institutions will move towards making those well-known repeated tasks much more automated, such as searches on external websites or data providers. This way you aren’t adding “headcount”, you are enhancing your analysts and allowing them to focus on gathering more information and making better decisions – overall making them more efficient.
- Service-Based Solutions (Cloud): Cost and risk are perennial concerns for executives and managers, especially when manpower and money are committed to on-premise deployments of enterprise software with no guarantee of success. In addition, it is a time-consuming process to on-board a new system or upgrade an existing system to keep up with the latest regulatory changes. Therefore, the future technology will be service-based cloud solutions, which are mainly Software as a Service — a software licensing and delivery model in which software is licensed on a subscription basis and is centrally hosted by vendor. Service-based solutions will also allow faster on-boarding of a system for a financial institution and will reduce overall cost, allowing compliance teams to focus more on the things they care about most, which is compliance. (The worldwide public cloud services market is projected to grow 18% in 2017 to total $246.8 billion, up from $209.2 billion in 2016, per Gartner, Inc.)
This change in trend will surely put additional burden on both financial intuitions and technology vendors. On one side, technology vendors will need to evolve to accommodate these upcoming demands, and on the other side, financial institution will have the biggest challenge to convince regulators in addition to train their staff and update policies. Like any other change, this phase will also take its own leap time and will require a few years to get to its mature stage.
It will be interesting to hear your standpoint on what other trends you are seeing in this space and the challenges and effects we should anticipate while the industry will move forward to acclimatize these new trends.