This article was written by Gregory Roberts for Bloomberg BNA.
Move over fintech, here comes regtech.
Entrepreneurs and innovators are invading the complex and often arcane world of regulatory compliance with some of the same technologies that have disrupted the core operations of the financial services industry.
Machine learning, biometrics and the interpretation of social media and other unstructured data show promise as potential compliance tools for regulatory obligations that have mushroomed in the aftermath of the financial crisis, along with the investment and costs associated with them.
Regtech is “the use of new technologies to solve regulatory and compliance requirements more effectively and efficiently,” as the Institute for International Finance (IIF), a research-oriented trade association in Washington, put it in a March report.
“I would define it as technological advancement that assists those focused on compliance and regulatory-related activities in their professions,” Kari Larsen, counsel at Reed Smith LLP in New York, and formerly in the Enforcement Division of the Commodity Futures Trading Commission, told Bloomberg BNA. “So making it easier, swifter, more complete, more efficient to monitor compliance and regulatory obligations.”
“For many years post-crisis, the only growing area of personnel, of hiring, in banks was in compliance,” Andres Portilla, IIF’s managing director for regulatory affairs, told Bloomberg BNA.
“Banks are spending enormous amounts of money in this area,” Portilla said.
Tech entrepreneurs and innovators moving into financial services initially focused more on what Portilla calls revenue sources, such as lending or money transmission, rather than on costs and ways to cut them, he said. Beyond that, a reason for an aversion to regulatory-related initiatives is that the field is dauntingly arcane and populated by seasoned experts, he said: “These are complex technical issues.”
Call for Input
In November 2015, the Financial Conduct Authority, a key regulatory agency in the U.K., issued a “call for input” on developing regtech. The IIF formed a regtech working group in response.
Although authorities in the U.K. may be ahead of their American counterparts in promoting regtech, the sector is growing in the U.S., Portilla said.
That’s part of the zeitgeist in financial services, Bart van Liebergen, IIF’s associate policy adviser for regulatory affairs and also a member of the working group, told Bloomberg BNA.
“It’s not a coincidence that we are seeing things more in regtech and in fintech at the same time,” he said. “We are seeing improvements in technology across the board — and we are seeing them used, on the one hand, in new funding models, and on the other hand, ventures are saying, ‘Hey, we can use this in compliance as well.’
“Where we think regtech stands out is that the new technologies allow banks, insurance companies and asset managers to automate parts of the compliance process they previously were not able to automate,” he said.
To Portilla, the question is, “How can those emerging technologies that are being used in fintech — how can they be leveraged to address the issue of compliance or regulatory reporting?”
The answers are still evolving, regarding which technologies to leverage and which compliance issues to address.
Tech Tools and Tasks
The applicable technologies include machine learning and artificial intelligence, biometrics, the interpretation of unstructured data such as e-mails and Facebook posts, and the use of application programming interfaces (APIs). Those tools can be brought to bear on such areas as aggregating big data, modeling risk for bank stress-testing, monitoring of capital-requirement compliance, updating compliance manuals, improving anti-money laundering and know-your-customer (KYC) programs and preventing fraud and in-house violations.
IBM’s chess-playing computer Deep Blue, which defeated world champion Garry Kasparov in a 1997 match, represents a classic example of machine learning. Deep Blue would record the positions of the pieces on a chess board before its move and, based on an information base of past games stored in its memory, chose a move to make. If that move backfired, Deep Blue would register the outcome and avoid a repeat when presented with the same situation in a future game.
Deep Blue’s descendant, the Jeopardy!-playing computer Watson, could be applied readily to regtech, IBM worldwide business development leader Neil Sahota said June 7 at a conference in Washington. Watson is engineered for machine learning and also for understanding ordinary language and interacting with the humans who produce it.
In a program with the University of Toronto, Sahota told Bloomberg BNA in a telephone interview, Watson is assisting lawyers by searching records of laws and court cases to determine which ones apply to a current case. The computer could perform a similar function with regulations, he said.
“You could teach Watson what the regulations actually are,” Sahota said. “They change in real time, and Watson could keep abreast of that. You could ask Watson to do a verification check to see if you are out of compliance.”
Mitek, a San Diego company that pioneered mobile photo capture of checks for bank deposits, and claims more than 5,000 banks as customers, has ventured into KYC compliance with features that allow consumers to photograph their driver licenses and other documents with their mobile phones when opening accounts.
“It became apparent that we could expand our product line to supply a growing market need for identity verification,” Sarah Clark, general manager of identity at Mitek, told Bloomberg BNA. “It helps banks and other financial services companies meet the customer identification program — aka, know your customer — requirements when it comes to onboarding customers,” she said.
Traditionally, a customer seeking to open a bank account would visit a branch in person and display a driver license or other form of identification. Now, Clark said, customers can take care of that step remotely, via smartphone. The bank can buttress the process by asking the customer to take a selfie, which Mitek’s Mobile Verify product compares with the photo on the license.
Mobile Verify has been adopted by major financial services companies for KYC, and they have not reported any problems with regulators. Mitek’s photo technology also can be used to capture other documents needed for financial transactions, such as applying for a loan, Clark said.
Regtech applications can bring in information from social media or other outside sources to buoy the anti-money laundering programs of financial institutions by fleshing out the profile of a customer whose activity has aroused interest, Larsen said.
In her experience, she said, regulators are receptive to regtech innovations.
“They’ve been open to meeting with clients who are using new technology, and to demonstrations of it,” Larsen said. “They tend to remain a little agnostic on the methods. I’ve yet to meet any real resistance.”
Ayasdi, a five-year-old artificial intelligence company in Menlo Park, Calif., is pushing both the reg and the tech sides of regtech, chief marketing officer Daniel Druker told Bloomberg BNA.
Ayasdi uses computers to mine data and crunch numbers far faster and more efficiently than humans can, streamlining risk modeling, stress testing and other compliance functions for large banks and other clients, Druker said.
“One of the things that were working with the banks on is, how do we bring the regulators into the process?” he said. “Now, for the first time, there’s time to do it,” he said.
“Before, they were always doing the mad scramble at the last second,” Druker added. “Now there’s time, and transparency, and there’s a chance for both sides to work together.”