As we wrote in July, financial institutions are increasingly using robotic process automation (RPA) alongside human operators to perform high-volume and repetitive tasks. This is allowing banks to shift human resources to more productive roles, such as analytics and due diligence, while accelerating processing and eliminating error-prone manual processes.
Regulatory compliance is one area where RPA is gaining traction quickly. Specifically, financial institutions are exploring the possibilities of using RPA in the area of anti-money laundering (AML) compliance where RPA can be used for both streamlining customer due diligence and transaction monitoring processes. Large banks typically have hundreds of employees assigned to these tedious tasks, cross-referencing customer data and poring over an endless and growing stream of transactions, looking for anomalies, such as, say, a customer who suddenly starts sending wire transfers to a foreign country.
It’s easy to imagine the difficulty of checking and cross-checking columns and numbers in a spreadsheet all day while maintaining a consistently sharp focus. Human and manual errors are both frequent and inevitable. However, such omissions are not permissible, despite the apparent difficulties, and AML fines can be in the millions of dollars for some of the larger institutions.
Banks are thus increasingly looking to automation to improve the efficiency of the transaction monitoring process, reduce the likelihood of errors and avoid penalties. In our experience, robots deployed alongside human operators in the transaction monitoring process are particularly well-suited for spotting anomalies and flagging them for the attention of their human counterparts, who are, in turn, better-suited for the more analytical work of research, analytics and due diligence.
If a financial institution is looking for ways to reduce compliance costs and increase accuracy and efficiency, the two natural places to consider deploying RPA are
repetitive or manually intensive processes, such as cross-checking customer and prospective customer names in web-based search tools or databases of bad actors, and
customer due diligence where RPA can be deployed for simple data search tasks. Human capital is much more valuable for the performing more complex and interpretive challenges.
We recommend a phased approach in designing and deploying RPA for AML compliance:
- Establish a governance structure for the implementation and use of the RPA tool
- Select the RPA tool that meets your business objectives
- Begin with a prototype, perhaps a single process or work group, and document the results. Given the complexities and upstream process challenges still present in the legacy systems operating at many financial institutions, this “shake down” approach will surface any compatibility challenges unique to your organization so you can address them.
- Define the steps in the process in detail and indicate what steps will be automated and how the automation tool is going to work. Transparency is important. Transaction monitoring is not an area where you want to have surprises. One advantage to consider is that RPA is simple and clear. Conversely, other technology (such as artificial intelligence) tends to be more of a “black box,” driven by algorithms that are often difficult to perceive and understand. With RPA, it is easy to see what is going on behind the scenes.
The good news is that regulators are encouraging “responsible innovation” in this area. As long as your financial institution plans carefully, does its homework, and is able to demonstrate governance acceptable to regulators, there is a high probability of success.
Ultimately, the success of RPA for AML transaction monitoring will come down to how well institutions are able to integrate the technology with their existing systems and how well they retrain and deploy the human capital freed up by robotic assistance. Defining clear strategic goals and documenting results is the best way to avoid automation for automation’s sake and reap the promise of RPA.
