Freeing high-powered AML analysts from rote tasks

Human analysis will always be required in assessing potential money-laundering risk. But automating routine activities allows analysts to focus on what they do best – risk assessment.

Financial institutions (FIs) are doing all they can to bolster Anti-Money Laundering (AML) programs and controls in the face of new rules. Yet regulators are still finding more non-compliance issues than ever, resulting in painful fines, reputational damage, and even, in some cases, personal risk exposure for AML officers. In response, many FIs have been hiring up – sometimes bringing on thousands more AML analysts. In fact, we’ve even seen some FIs increase their AML staff by as much as 500%, causing a headhunting free-for-all as firms desperately search for the right experts.

Yet rather than solving compliance issues, this rapid expansion is causing a host of new problems, including:

• Burgeoning costs: Thanks to battalions of new analysts, the cost of compliance has skyrocketed. To control their AML budgets, companies sometimes tinker with their transaction monitoring systems – exposing their firms to new risk in the process
• Talent disparities: The bigger the team of analysts, the more likely their expertise will vary. The highly experienced will deliver superior results. But those less skilled may miss critical indicators, opening up their institutions to risk
• More human error: When AML investigative processes rely heavily on an ever-growing number of humans, the odds of human error occurring increase

Does the digital revolution have a role to play in mitigating these issues? No question, there have been real advances in technologies and artificial intelligence (AI) over the past decade. But these technologies still can’t beat people for critical thinking—especially when it comes to making decisions about risk and compliance that involve complex variables and that demand context and instinct.

The fact is, no algorithm will ever take the place of a seasoned AML expert’s risk assessment. Even the most complex algorithms can’t compete with someone who has spent a lifetime in the field. Experienced AML analysts see hundreds of case investigations in the course of their work. That gives them an almost unconscious knack for detecting types of risk – and an understanding of human behavior – that helps them recognize anomalies. What these technologies can do, however, is take over some routine activities so experts can focus on what they do best – analysis.

Understanding how to get the most out of these technologies requires a hard look at how analysts spend their time. While each FI takes a different approach to its financial crime program, there is some common ground. For example, most AML programs consist of KYC, onboarding, monitoring, and investigation processes, typically involving the following steps:

• Data collection: Analysts generally follow defined procedures that involve collecting all available information about the subject of an investigation, including bad press, watchlist matches, and network risks. Commonly, they conduct searches using Worldcheck, Bridger Insight, Factiva, Google, and other sources. This process takes up to 75% of an analyst’s time
• Organization and data entry: Once analysts have the information in hand, they must enter it into the system – an important step in the audit trail. In many cases this entails uploading documents to a case management tool, copying and pasting information from a browser, or preparing extensive notes. Analysts spend 15% of their time doing this
• Risk assessment and report writing: The final step, which calls for muscular AML expertise, involves synthesizing the collected information, assessing risk factors, determining appropriate action, and preparing a report. Analysts spend 10% of their time on this step

A generalized time study helps identify inefficiencies and opportunities, specifically in the first two steps. Gathering and entering data are repetitive exercises, and technologies already exist for automating these functions – technologies that let FIs define their own due-diligence protocols for negative news searching, network analysis, and so on.

This automation drives efficiency and cost reduction, but it can’t help with assessing risks and writing reports. With aggregated data at their fingertips, liberated analysts do what they excel at – risk analysis. That takes the power of the human brain. And with an estimated 90% of its time freed up, the human brain can work wonders.