Building Sustainable AML Practices With Intelligent Automation

Ciarán Daly

March 11, 2019

5 Min Read

by Kyle Hoback and James Lawson

LONDON - Money laundering is an ongoing challenge for banks and financial institutions. The fight against it traces its roots to the 1980s “War on Drugs” in the United States and the first EU directive on Anti–Money Laundering (AML) efforts in 1990.

The related pressures, complexities, and costs have escalated dramatically since, and the sector’s response has been to create ever-larger operational teams performing slow, manual work. Now it is clear that a reliance on manual work is not a solution. In fact, it is a root cause of many of the challenges associated with AML.

One answer is Intelligent Automation, which eliminates the problems of a labour-intensive approach while still keeping people involved in important decisions.

Current methods can’t cope

Traditional AML operations are time-consuming and unfulfilling. There are a high number of errors. Clients are left unsatisfied by repeated call-backs and long onboarding processes. AML capabilities are inadequate, with huge numbers of false positives and questionable risk assessments. With each failure, the threat of further fines looms large.

Since the global financial crisis that started in 2008, Aurexia Institute reports there has been over USD $450 billion worth of compliance penalties worldwide. Regulations are also becoming more complex to comply with; since 2016 alone, compliance teams have faced over 51,000 regulatory changes. The typical “bulge bracket” institution is spending over USD $1 billion annually on compliance.

Regulators encouraging AI innovation

For decades, the response to more crime was to hire more people to detect and fight it. This is not sustainable nor efficient, neither in terms of costs nor time. But how else are financial institutions supposed to strive toward eliminating money laundering?

The true solution is to embrace Intelligent Automation, which blends the best of automation technologies, such as AI, with the expertise of their personnel teams to improve productivity and compliance while reducing cost.

It’s no surprise that regulators are encouraging financial institutions to embrace innovative approaches. Federal Banking Agencies and the Treasury’s FinCEN issued a joint statement in December 2018, specifically welcoming artificial intelligence to “strengthen compliance approaches… enhance transaction monitoring system… [and] … maximize utilization of banks’ BSA / AML compliance resources.”

Related: Weak Supervision - How A New Technique In Machine Learning Makes AI Easy For Everyone

Now there’s a pre-built solution

WorkFusion, the leading Intelligent Automation platform, has gone one step further. For common manually intensive processes, it offers pre-built solutions. These specially pre-trained “expert bots” enable a bank to solve problems much more quickly, like the analysis of adverse media, source of wealth and ultimate beneficial ownership during an Enhanced Due Diligence.

Complex processes like these were previously thought too challenging to automate. Now a bank can onboard a robot in weeks, as quick as it would take to onboard a new employee from a competitor’s AML team. Simultaneously, that robot will learn, improving over time as it receives more training.

In addition, WorkFusion’s Smart Process Automation platform provides the flexibility to tackle new AML processes, and meet specific internal requirements — just like they might have done historically with new analysts.

AI can be tested in a 2–4 weeks, and in production within 8–12, with same-year ROI. These AML processes, which combine learning robots, analytics and workflows, quickly see a 40–90% productivity boost.

Improvements like these enable banks to dramatically improve their operations, and leave people free to focus on more valuable activities, quality control and continuous improvement.

Example: Expediting Negative News Screening

Let’s consider one AML application (among many possibilities) in more detail: searching for negative news. In this key AML process, analysts must review individuals and companies during onboarding, and conduct periodic reviews based on risk profile.

In both cases, they search news sources, including Factiva, Google and others, and use only their judgment to interpret hundreds of results. They do this in search of anything negative which suggests the customer or client is a risk — triggering further investigation, if necessary.

Analysts using WorkFusion save huge amounts of time. News is automatically downloaded and reviewed by the package’s learning bots, which assess whether it is relevant to the entity in question.

These bots then analyze the content for PEPs (Politically Exposed Persons), Sanctioned Countries and most importantly, risk factors presented within the news report itself. The bots can determine whether the bank should be concerned, then produce a summary of findings.

Ultimately, the analyst reviews the work before the bot writes a final report and updates core systems. All in all, this achieves a reduction in manual work of up to 80%, while increasing coverage and scalability.

It’s obvious that current approaches to fighting money laundering and other financial crimes must be updated to meet the growing worldwide threat. Institutions cannot hire enough people to ward off countless bad actors bent on outsmarting outdated security methods.

The more successful solution is installing continuously learning bots to serve as a broad first line of defense, with a sustainable number of subject-matter experts providing judgment and escalating responses as needed. This approach saves time, increases employee satisfaction, improves compliance, and — perhaps most importantly — protects the integrity of both the institution and its customers.

rete-180605-0949.jpg

Keep up with the ever-evolving AI landscape
Unlock exclusive AI content by subscribing to our newsletter!!

You May Also Like