ML and automation drive efficiencies in anti-financial crime efforts

While machine learning (ML) and automation are helping to fight financial crimes like fraud and money laundering, there's still plenty of room for improvement. Since the onset of the pandemic, anecdotally, there's been "a huge uptick" in first-party financial fraud, or fraud perpetrated not under false identities, said Sandip Nayak, chief strategy and artificial intelligence officer at Linear Financial Technologies , provider of a digital account origination platform. Only 1% or less of the $2 trillion to $4 trillion laundered each year through the global financial system is detected, according to estimates from the United Nations and anti-financial crimes software developers like Nasdaq . That success rate is "absolutely appalling," said Andrew Davies, vice president of global market strategy for financial crime risk management at core provider Fiserv . Limitations of rules-based systems Part of the difficulty in anti-financial crime efforts is that financial institutions have largely used rules-based systems, which try to identify fraud based on prior fraud instances, explained Jason Chorlins, principal banking practice leader at accounting and business advisory firm Kaufman Rossin . These systems tend to return large amounts of false-positive alerts, he said, bogging down anti-money laundering (AML) and anti-fraud resources. Rules-based engines use identity data elements such as name, address, phone number and device, which are not predictive, Nayak noted. The ripple effect of false positives also ties up time and resources. Those alerts typically come after significant instances of financial crime. Fraud schemes start out in small numbers before building up to where they trigger the flags that rules-based monitoring systems rely on, Chorlins told Bank Automation News . Banks and FIs "can have the best rules in the world, but if they're not tailored to identify the specific fraud typologies that are out there, then more likely than not, they're going to go undetected — they're going to go under the radar," Chorlins added. 3 factors for smarter detection Deep-learning algorithms. Chorlins and Nayak told BAN that better anti-financial crimes systems address potential criminal behaviors within the population, rather than searching for data elements. Advanced detection systems use deep-learning algorithms, a type of ML, they said. Alternative data. Detection systems can also use alternative data to improve outcomes, Nayak said. Linear's anti-fraud platform , which launched in November, leverages alternative data streams — and with ML, the more contextual data, the better. That might include things like additional background information on the payer and recipient of a transaction. Auto-calibration and customization. A third feature of advanced detection systems also falls in the ML category: auto-calibration and auto-customization. Fraudsters evolve and the nature of fraud can be very channel-specific, Nayak said, such as whether it's occurring at a physical bank branch or online, and an advanced detection system must also evolve if it is to be predictive. ML technologies are examining financial transactions and trying to find exceptions from usual behaviors, said Chorlins. Advanced detection systems also are honing in on more specific populations, targeting narrower "peer groupings" rather than the broad business groupings that a rules-based system would address. Automation is playing a role particularly in managing mundane tasks like sorting through alerts and red flags. Robotic process automation (RPA) can take routine tasks like manual searches for additional information on those involved in a flagged transaction and automate them, Chorlins told BAN . Accelerating change As improved fraud and financial crime detection systems using ML proliferate, more banks and FIs are adopting them. In the past two years, "we're seeing more widespread adoption of these different types of technologies," Chorlins said. And the spike in remote work since the onset of the pandemic has accelerated adoption. "Almost every single compliance person is saying, 'Well, my technology better have some sort of machine learning in it — I don't want to be behind," added Chorlins. Bank Automation Summit, taking place March 1-2 in Charlotte, N.C., is the first and only event to focus solely on automation in banking. The event will feature the brightest minds from across financial services on intelligent automation strategies and deployment. Learn more and register for Bank Automation Summit 2022 .

ML and automation drive efficiencies in anti-financial crime efforts

FIs move to advanced detection systems over older rules-based systems

Truist CIO Bryce Elliott joins Bank Automation Summit for fireside chat

5 questions with … JPMorgan Chase Consumer and Community Bank CIO Gill Haus

Related Posts

Lloyds keeping ‘eye on’ AI insurance trend

TD commits up to $25M to Cohere partnership to accelerate AI adoption

C-Suite alignment is ‘most important piece’ of AI puzzle

5 questions with … JPMorgan Chase Consumer and Community Bank CIO Gill Haus

EMERGING FINTECH DIRECTORY

FinAi Podcast

SPONSORED

Build an Antifragile Strategy to Outperform the Market

How AI and Product Experts Turn Fuzzy Requirements Into Focused Dev-ready Roadmaps

Is Your Technology Supplier There for You?

Connect

Welcome Back!

Retrieve your password

Continue Reading with FinAi News Premium - Less than $2/Day

ML and automation drive efficiencies in anti-financial crime efforts

FIs move to advanced detection systems over older rules-based systems

Truist CIO Bryce Elliott joins Bank Automation Summit for fireside chat

5 questions with … JPMorgan Chase Consumer and Community Bank CIO Gill Haus

Related Posts

FinAi Podcast

Connect

Welcome Back!

Retrieve your password

Unlock This Article

Continue Reading with FinAi News Premium - Less than $2/Day