Robots and Automation in AML: The Future is Now

Topics: Artificial Intelligence, Financial Crime, Financial Fraud & Anti-Money Laundering, Fraud, Government, Risk Management


MINNEAPOLIS, Minn. — Sci-fi has always tantalized audiences with the promise of the forthcoming days when robots will simplify our lives. Whether we are talking about Rosie the robot maid from the cartoon The Jetsons, HAL (Heuristically-programmed Algorithmic Computer) from Stanley Kubricks’ 2001: A Space Odyssey, or even R2D2 from Star Wars, the concept of robots as partners to humans has always been a near-distant, future concept.

Yet now we have self-driving cars and friendly virtual assistants in our phones named “Siri” or “Alexa” that can help us in ways unheard of even five years ago.

But what about the banking industry? Surely robotics hasn’t evolved to the point where banks can simply put robots to work at large-scale tasks? We must be at least a lifetime away from the kind of progressive technology that would make humans obsolete, aren’t we? Not quite.

While humans are still very much a part of a bank’s overall fight against financial fraud, the future is now. Robotics and automation are here and will help with operational demands to advance the digital and technology agendas related to anti-money laundering (AML) processes in financial institutions.

But the results are only as good as the data we put into it, argued experts at a recent learning session hosted by the Twin Cities chapter of the Association of Certified Anti-Money Laundering Specialists (ACAMS) in partnership with Thomson Reuters. The group discussed the increasing industry trend and operational demand of robotics process automation (RPA), machine learning, and artificial intelligence (AI).

Panelists included Holly Sais Phillippi, partner director for Governance, Risk & Compliance at Refinitiv; Debra Geister, Business Strategy & Operations for Section 2 Financial Intelligence Solutions; Andrew Bingenheimer, SVP in Financial Crimes Development at U.S. Bank; David Berglund, SVP and AI leader at U.S. Bank; and Catherine Banks, market specialist in Compliance & Risk for Refinitiv.

Industry Driven Trends

Trends drive industries; and nothing drives a trend more than lost profits. U.S. financial institutions are spending more than $8 billion dollars annually in AML compliance — but seizing less than 0.2% of laundered money, according to the panelists. When it comes to transaction monitoring, financial institutions are generating an average of 90% false positives. A false positive is essentially a false lead on a suspicious transaction that eventually turns out to be nothing.

Further compounding the push towards AML technology solutions are the fines levied against banks themselves. “Regulatory agencies are fining financial institutions for aiding and abetting terrorism, organized crime, and sanctioned persons, entities, and nation-states,” Geister said.


ACAMS Panelists (l to r) Andrew Bingenheimer, David Berglund, and Debra Geister listen in as Catherine Banks answers an audience question about robotics technology.

AML enforcement by regulators is a Top-5 loss event for financial institutions, yet the fine is only half the story, she explained. Other areas of loss include:

  •  Stock share values decline 5.5% the day fines are announced;
  •  Cease & desist orders result in loss of new programs, vendors, and business plans; and
  •  Remediation costs over the first 18 months are 12-times greater than the fine itself.

Hence, financial institutions are getting hit hard by regulators and are spending more for less satisfactory results. The increased, overall compliance burden from the regulators — such as the Ultimate Beneficial Ownership (UBO) reporting requirements — further drive the move towards technologies like RPA.

RPA and Machine Learning Defined

So, what is RPA? The term refers to software that can be programmed to do more basic tasks across applications. Its purpose is to reduce the burden of simple, repetitive tasks. For instance, it can be used for rote functions like data collection and can help automate Know Your Customer (KYC) processes. This creates efficiencies, minimizes investigator time and ultimately reduces costs.

Machine learning aims to increase accuracy. It features a self “learning” algorithm which allows for learning from the data supplied. Simply put, it is an umbrella term for simulated intelligence in machines that are programmed to “think” like a human. These machines can then perform the more advance actions such as learning and rationalizing and take the action that has the best chance of achieving a specific goal in the most effective and efficient way.

This is important because a big concern for banks involves so-called hybrid threat organizations, which perform transactions too complex and sophisticated to be discovered by routine AML monitoring, according to Geister. In fact, most financial institutions do not train their financial intelligence units in transnational security issues, terrorism, organized crime, and counter-threat finance tradecraft needed to identify sophisticated operations. Using RPA or machine learning is one way to improve the quality of the alerts and quality of suspicious transaction reporting, which naturally improves the overall quality of investigationsTo do that, however, we need to better understand and target the threat network and follow the money.

As we know, the biggest question is no longer, “will we ever be able to use RPA in AML processes?” The question is, “how can we best use RPA and machine learning in improving overall efficiencies?”

When deciding what to automate, this is a question that each financial institution will need to tailor to their specific needs. Can RPA quantifiably reduce the amount of time it takes to report a client risk assessment alert? Will the results demonstrate a true benefit of standardization, quality, and consistency? Did a financial institution’s pivot towards RPA, automation, and AI increase risk coverage, decrease false positives, and lower compliance costs, all while focusing on informed money laundering and terrorism finance typologies?

Only time will tell, but the future is now and judging by the increase demand for efficiency, we won’t have to wait long.

Robotics and machine learning are coming to financial institutions. As a risk and compliance professional, stay up-to-date on the latest technology-fueled AML trends by following the Legal Executive Institute’s Risk & Compliance page.