Software for detecting financial crime is in high demand. Here’s why.
Investments in advanced software that can be used to detect money laundering are booming. Over the last year, banks and venture capitalists have poured tens of millions of dollars into the sector.
On Wednesday, Hummingbird, a 3-year-old company that helps banks and fintechs with anti-money-laundering compliance, announced an $8.2 million Series A round led by Flourish Ventures. Last month, Nice Actimize, long the leader in financial crime detection software, announced it had agreed to acquire a competitor, Guardian Analytics.
An AML software startup, TookiTaki, raised $11.7 million in November. Salv, founded by former Transferwise employees, raised $2 million in December, and Elliptic, which offers software to detect crime involving cryptocurrency transactions, received $23 million last September.
At the same time, banks are going public with AML modernization projects, something they’ve rarely done in the past. In June, Banco Santander announced that it is deploying software from ThetaRay. Last year, HSBC said it had bought financial crime detection software from Quantexa, and BMO Financial Group recently told American Banker it spent $1 million to develop a new AML system internally.
Why is this happening now? There are at least four reasons.
To avoid steep AML fines
Last year, regulators imposed $8 billion in AML fines against banks globally, including 12 of the 50 largest institutions. The penalties have continued to accumulate in 2020.
In July, New York State regulators fined Deutsche Bank $150 million for failing to properly monitor its relationship with the convicted sex offender Jeffrey Epstein. The regulators said the bank processed hundreds of transactions for the late financier, including payments to Russian models and $800,000 in suspicious cash withdrawals.
In June, Commerzbank in Germany was fined $47 million by U.K. regulators for anti-money-laundering lapses. In April, Westpac in Australia said it had set aside $570 million for a potential fine stemming from money laundering and allegations of facilitating child exploitation in the Philippines.
Individual bankers are also facing risks. In March, the Financial Crimes Enforcement Network fined former U.S. Bank executive Michael LaFontaine $450,000 for his failure to prevent violations of the Bank Secrecy Act and anti-money-laundering rules.
To keep up with sophisticated criminals
One impetus for banks is the need to keep pace with criminals who constantly hone their skills.
“You've got more sophisticated financial criminals who are using technology to fly beneath the radar, to conduct more complex and faster transactions, and they're taking advantage of digital experiences and faster payment rails,” said Charles Subrt, a senior analyst at Aite Group who specializes in fraud and AML.
The United Nations estimates that $2 trillion is laundered every year, and banks help catch less than 1% of that crime.
The older systems many banks use are often blamed for generating high numbers of false positives, which create time-consuming and fruitless investigations for human compliance teams without catching criminals. Money launderers know the rules and know how to structure their transactions in ways that do not trigger the software.
“The detection models have historically been simplistic and rule-based,” said Meera Das, managing director of AML modeling and machine learning at BMO Financial Group.
The need to be able to explain to bank regulators how the software works has been a large hindrance to using more complex models, she said. But this is changing.
“Everyone's gotten wiser in terms of how financial crime occurs,” Das said.
Newer technologies like robotic process automation and machine learning that can be used to quickly pull background information on parties to a transaction “are becoming more the norm and becoming more accepted,” Subrt said. “So you're seeing further and further investment in that.”
Regulators are embracing modern tech
Kabir Kumar, director of Flourish Ventures, which led Hummingbird’s Wednesday funding round, pointed out that regulators themselves are becoming more tech-savvy.
“I think what's interesting about this moment is the path that we've been on to transform regulation and to digitize all these aspects of compliance has accelerated,” he said.
Regulators are pushing banks to use risk-based approaches to combating financial crime, fraud and money laundering, Subrt noted.
“They’re really focusing on outcomes and less on a check-the-box type of process,” he said. “And they're pushing and encouraging innovation within the financial crime space; they appreciate that only with innovation can you really, truly combat financial crime. You're seeing a number of agencies like Fincen and the FDIC set up innovation hubs to foster that joint collaboration with the private sector.”
Consumers’ changing behaviors are spurring adaptation
The pandemic has rapidly changed consumers' and businesses’ behavior, with more online shopping and fewer in-person purchases, said Joe Robinson, CEO and co-founder of Hummingbird.
As a result, “Banks and fintechs are rethinking some of the models they use to monitor for AML,” Robinson said.
James Heinzman, executive vice president, financial services solutions for ThetaRay, said his mother recently got an iPhone and began doing all of her banking digitally.
“She had never ever done an online transaction before,” he said. “Now that's all she's doing, and she's not going back.”
This kind of behavioral change makes it harder for algorithms to define what is normal and what's not, Heinzman said. (Machine learning-based systems like ThetaRay's, he said, have no predefined models and rules; how they define normal activity is based on the current context.)
The pandemic has also changed some bad actors’ behavior, Heinzman observed. They’re no longer flying suitcases of cash around.
“A lot of the things that they were doing in the past, they simply can't do,” he said. “It's also creating a lot of new and evolving schemes that the banks have to try to figure out how to catch and identify.”
Bankers may still have qualms about using some newer technology for financial crime detection, especially any form of artificial intelligence, for fear of what regulators will think.
If the software uncovers a lapse that has existed for years, will the bank be prosecuted? Will field examiners understand the new software or accuse the bank of using “black box” technology?
Regulators have expressed a willingness to allow banks to try such software, but bankers still worry that they may hear criticism or suffer other repercussions.
“I would say we're in a journey,” Subrt said of such concerns.
Das said regulators have been very open to the AML project she has been leading at BMO Financial.
“We've never had any pushback, as long as we've explained how we've approached a problem, what we've done that’s different from what was there before and why we're confident that this is a better solution,” she said. “And part of this is also taking the time with the regulator to ensure that they are comfortable, and seeing if they do have any concerns.”
Rather than ripping out and replacing their existing AML systems, some banks are augmenting them. For instance, BBVA uses robotic process automation to help with data gathering, which frees up investigators to focus more on analysis and decision-making.