By Anuj Khurana, Co-Founder & CEO, Anaptyss
Complying with anti-money laundering (AML) regulations has been challenging for financial institutions as they continue to scale up their efforts and resources. Amid heightened scrutiny by regulators and the proliferation of financial crimes, tracking suspicious transactions, conducting KYC, and screening sanctioned entities/individuals have become inordinately resource-intensive and risky with razor-thin error margins.
Industry figures estimate upwards of US$ 200 billion as the annual cost of meeting
AML compliance in 2021, and this number continues to grow exorbitantly. Any violations can entail additional costs via legal penalties, revoked licenses, and loss of goodwill.
Consequently, banks now look for technological intervention to “operationalize” their financial crime investigation units (FCUs) to offset the hefty costs and risks. Technologies like machine learning and robotic process automation (RPA) have emerged conspicuously with the promise of increased efficiency and accuracy in combating money laundering. For instance, machine learning can demonstrably reduce false positives and expand algorithmic intelligence to large-scale surveillance and customer due diligence (CDD).
Likewise, the RPA realm can automate rules-driven processes in the AML and KYC value chain, dramatically reducing manual intervention and increasing the throughput.
Sea of automation opportunities
The application scope of intelligent process automation tools aka software bots spans four areas: KYC, watchlist screening, transaction monitoring, and customer offboarding.
KYC processes consume vast resources in manual activities like feeding customer information into the CRM, validating customer data, maintaining financial data such as credit scores, etc. These tasks require repetitive steps such as uploading a document, merging information in a consistent format, controlling document versions, etc.
Programmable bots offer a fast and precise method to automate KYC processes using techniques like data extraction, document indexing, copying and insertion, and analyses. They can be explicitly coded to emulate logical rules, freeing up human capacity for strategic tasks.
Watchlist screening is critical to customer due diligence, wherein banks need to continuously comb a variety of databases, including OFAC lists, public records, news portals, and social media to screen sanctioned entities, politically exposed persons (PEPs), and blocked individuals.
With an estimated 24,000+ sanctions worldwide, the effort and risks remain at an all-time high. RPA, combined with machine learning and natural language processing, can equip banks with a powerful capability that can parse authenticated public databases to enable real-time surveillance of sanctioned entities at a global scale.
Among all, AML transaction monitoring demands the most time and effort from FCUs. With over 90% of false alerts generated by traditional transaction monitoring systems, tier-1 analysts are compelled to make redundant efforts, inflating the time
and costs.
RPA backed with machine learning can detect false positives and false negatives with higher accuracy based on the evolutionary learning capability. Trained using supervised and unsupervised approaches, these automated AML systems can mechanize alert generation, reviews, hand-offs, and analysis tasks to a great extent, reducing manual, repetitive efforts.
Key benefits of machine-learning-based AML solutions include automatic risk categorisation, untampered alerts, and tamperproof audit trails for compliance.
Offboarding & closure are other areas concerning high-risk customers. Banks need to conduct enhanced due diligence (EDD), which involves communication, recordkeeping, status updates, etc. While financial institutions decide the account offboarding status, they can leverage automation tools to speed up EDD and closure efforts.
Considerations for RPA adoption
Several factors underpin the successful adoption of RPA tools, and these include – expected outcomes, choice of solution, process readiness, technological maturity, risk oversight, and cultural synergy.
Foremost, financial institutions need to evaluate their existing manpower costs for AML compliance, breakeven point, and functional/non-functional benefits before investing in an RPA tool. This will set the expectations right at the onset. Next, institutions need to consider the extent of “tailoring” and vendor support available to adapt the automation tool to their specific requirements. It’s a misnomer that automation is the best strategy for any process. The fact is that a poorly designed process can hamper automation efforts, resulting in wastage and undesired outcomes.
Therefore, process audits and reengineering are crucial to preparing it for RPA implementation. Another consideration is the organization’s technological state; a financial institution just starting with adopting digital technologies vis-à-vis an organisation with a matured technology setup will have vastly different approaches to adopting automation. For instance, interoperability and compatibility are critical factors a bank should assess before bringing an RPA tool into its AML ecosystem.
Additionally, risk oversight must be built within the AML compliance ecosystem.
Banks need to consider that the most robust systems elevate human ingenuity with
technology.
Thus, for adequate oversight, the automation rules need to undergo review by domain experts, and human specialists need to check critical/high-risk alerts and provide failover support to automation systems. Remember, bots can fail too, and when they do, there must be human oversight.
Culture-and-domain-first approach
Automation can drive a fundamental shift in operations, and its success depends on
an agile culture that supports change.
In tandem, domain-centric expertise provides the fulcrum to plot strategic elements,
from optimal consulting to guided implementation, to ensure the outcomes align with
the organization’s expectations. A realistic approach that continually hones the business context is imperative to driving the cost and efficiency benefits.
