Leveraging Data & Tech in Financial Crime Compliance

The pandemic has been an agent for accelerated change in operational and frontline functions. Financial crime compliance teams need to get in on the act.

Financial institutions in Asia have proven critical in supporting policymakers’ stimulus measures and other economic initiatives amid the Covid-19. In many cases, this has meant the adoption of new tools and technology to quickly integrate electronic KYC, digital signature collection, and online document submission into core bank processes to ensure customers could continue to access financial services remotely.

In this sense, the pandemic has been an agent for accelerated change. While many financial institutions have updated their KYC and onboarding processes earlier than they otherwise would have under normal circumstances, some have even fast-tracked procurement protocols – to more quickly enable seamless digital processes, even against the difficulty presented by remote working arrangements.

Further complicating this work, new patterns in criminal behaviour have emerged during the pandemic, placing enormous strain on fraud, AML, risk management and compliance teams responsible for ensuring appropriate controls are in place to prevent financial crime.

“Bad actors who were previously laundering money through hawala and other cash-based means have been increasingly working through the banking system digitally,” says Crime Stoppers Asia CEO Richard Carrick. “This is testing banks’ detection of financial crime at a time when their abilities to put in new transaction models are also being tested.”

Meanwhile, a shift in consumer spending patterns amid the pandemic has meant that existing systems for monitoring suspicious activity have needed recalibration. While a lot of banks have the capability to analyse standard transaction monitoring alerts remotely, staff members were quickly be overwhelmed by the sheer number of alerts they had to assess.

In many cases, banks simply adjusted their risk scoring models or thresholds to ensure they were not generating too many alerts, as additional resources and skilled personnel have been difficult to come by on short notice.

From a financial crime standpoint, financial services firms can ease some of these challenges by using newer technology to automate away inefficiency and by addressing data gaps that limit the ability of firms to screen false positives.

According to Matthew Field, AML Market Director for Asia Pacific at NICE Actimize, the difficulty is the amount of time it can take to review each alert with the same investigative eye and attention to detail that is needed. False positives were named the most significant challenge facing investigators at financial institutions, according to NICE Actimize research.

Indeed, each alert requires a significant amount of time and effort to investigate, estimated at somewhere between five and 30 minutes for a single low-level false positive. More complex alerts can sometimes take days.

“The key is to establish practices that can reduce the percentage of false positives to more manageable levels,” Field says. “The traditional technology solutions used in AML operations have struggled to scale to modern day challenges, which have only been accelerated by Covid-19.”

Technologies such as AI (artificial intelligence) and machine learning are already being used to enhance AML programmes at financial institutions. In November, the HKMA (Hong Kong Monetary Authority) surveyed of 196 institutions on their use of regtech, finding that about 33 percent were already using AI (including machine learning) in their financial crime programmes, most prominently in monitoring, screening and investigations.

Indeed, many banks in the region are hiring data scientists to develop models that can mimic human decision-making, but at higher speeds and with a higher degree of efficiency, as they seek to automate away a large part of the manual work typically done to identify an alert as a false positive.

According to Field, new AI-based approaches to tackle financial crime are already yielding positive results – lower false positives, identification of false negatives, and improvements in overall investigative efficiency and productivity.

Some financial institutions are also using machine learning to improve risk segmentation, alert scoring and prioritisation, behaviour monitoring, pattern recognition, anomaly detection, and document and text screening.

Meanwhile, robotic process automation is being used to achieve operational efficiencies in typically labour-intensive processes, including case investigations, data extraction and reporting.

Yet, many of these new tools – and the effectiveness of AML programmes – rely on a bank having in place a sound strategy for managing data, widely considered the foundation of any good AML programme.

In recent years, banks have begun to revisit their mechanisms and processes for handling data, introducing changes to break down data silos and simplify data architecture. They are also increasing the reusability of data by standardising data definitions and formats, and by normalising data models across products, business lines, and regions.

Meanwhile, software providers have centralised access to financial crime-critical data sources – i.e. new sources, rating agencies, corporate databases and watchlists – helping firms to save on time and costs involved with sourcing that data themselves. This data, when fed into case management systems and investigations can be used to provide third party monitoring and background checking.

Still, other data challenges persist. Differences in market practices, regulatory requirements, and product specificities make it difficult to unify data in its entirety. But, part of the problem is that many financial institutions are constrained by legacy systems and processes, unable to address broader bottlenecks and limitations.

With the right data infrastructure in place, leveraging on both internal and external sources, and complemented by technology that enables more efficient decision-making, financial institutions will be able to better manage false positives and caseloads, while also reducing the need for threshold adjustments, risk scoring model tweaks, and other stop-gap measures.

The pandemic has already been an agent for accelerated change across operational and frontline functions, resulting in efficiency gains for financial firms across the region. It is equally important for financial institutions to ensure their risk management and compliance teams get in on the act and leverage the technology that is available.

This article was adapted from a whitepaper jointly published by Regulation Asia and NICE Actimize. The full paper is available for download here.

 

To Top