Frank McKenna is Chief Strategist at Point Predictive, an AI solutions company providing predictive scores for auto, mortgage and fintech lending. He’s also the author of the widely read blog FrankonFraud, where he analyzes fraud trends as they happen globally and reports on industry impacts. An advocate for fraud managers, fraud data scientists and fraud fighters around the world, Frank has worked with over 200 banks, lenders and finance companies. He talked with Mike Sasaki, leader of the Mitek Systems global customer success team, about what he’s seeing with identity and fraud, and what he thinks comes next.
Mike: Hi Frank. What are the most significant changes you’ve been seeing over the past year or so?
Frank: We’re seeing a surge of social engineering scams. Many of these fraud schemes are pretty complex, such as forging documents to target stimulus or paycheck protection funds. On the e-commerce side, hiring professional social engineers to conduct refunding fraud; these are people who know exactly how to convince online retailers that received packages never arrived. We’re also seeing a surge in identity theft. Something like one in 50 consumers has already been victimized. Over the last 18 months it’s really been ramping up, and every day I’m seeing more and more.
On the positive side, we’re also seeing a huge surge of investment in fraud defense technology innovation. In one recent month alone, there were five unicorns in the area of fraud.
Mike: Let me ask you to switch your hat for a minute from industry analyst to chief strategist at Point Predictive. What are the technology innovations that have helped you most over the past 12 months?
Frank: Cloud computing from AWS has helped Point Predictive deliver analytics, machine learning and other modeling in a very efficient, rapid way. It used to take us two to three weeks to develop a model; now we do it in a day because of all the tools we have in the cloud. We can engineer new models and deliver solutions very quickly, and also score millions of transactions in seconds because we can spin up thousands of lambdas (essentially, serverless computers) in the cloud as needed.
But the real game changer for us has been robotics, and robotic process automation (RPA) in particular for identity. Every month Point Predictive analyzes data from two to ten million credit applications submitted to lenders and dealers. We discovered we could use RPA at night to comb through this data and identify patterns indicative of criminal rings—things like multiple applications using the same employer or home address. RPA has enabled us to supercharge the identification of clusters of suspicious behavior, increasing the number of alerts we send out by 200%. That’s really helping our customers take action against fraud rings, and it’s been a real game changer for us.
Now, combine this ability to deeply mine data with all the alternative data becoming available in the industry from social media, search tools and other sources. That’s becoming super-helpful for identifying and investigating fraud, and being proactive against synthetic identity fraud.
Mike: How do your customers view identity verification?
Frank: Well, two of our markets, auto finance and mortgage, aren’t known for being the most advanced with digital technologies. Still, every single lender we’re working with has a digital transformation program and is concerned about enabling consumers to take out loans online with minimal friction. Their biggest concern—and this is true of our other major market, fintech lending, as well—is that they don’t want a high rate of false positives (legitimate customers identified as suspicious) because they don’t want to be turning consumers away.
So even though we provide fraud defense services, we see our mission a little differently. If you’re really good at identifying the tiny percentage of transactions that are truly suspicious, then you’re really good at telling lenders: Here are the 97-98% of applications you can trust. That enables them to safely automate their lending process. Our role is to help ensure that consumers who are being truthful in their applications for credit flow through the process uninterrupted. Catching fraudsters is a byproduct.
Mike: What is your thinking about biometrics for identity verification and the concerns about potential bias?
Frank: I think biometrics, such as the facial and fingerprint recognition that have been made so popular and commonplace by Apple for one, are really transforming the way payments are made. And they’re going to transform how loans are made as well. I see them as keys for enabling a lot of very quick and secure lending.
Consumers, I think, are going to be very willing to use biometrics in this way. In fact, I think this is the wave of the future for lending—getting consumers to interact with you and provide key pieces of information so they can flow more freely through the process and also protect themselves. My view is that this is something people are really going to gravitate toward.
Regarding bias, it’s there, but with technology improving and data collection increasing, we’re going to quickly overcome it. We’ll be able to achieve better segmentation in the AI to really look at demographic groups so that we have the best recognition with no differences in performance from one group to another.
When are we going to see widespread adoption? Auto and mortgage lending, as I said, tend to be behind the curve in tech adoption, but they’re also among the markets with the most to gain since they have relatively high rates of fraud, and it’s moving online. So I expect biometrics to become part and parcel of the secure lending process in these markets, but it’s probably 18 months to two years out.
Mike: Which is more important for driving progress: consumer adoption or regulation?
Frank: Both, they’re intertwined. Consumer privacy groups seeing all this data being collected out there are raising the red flag, which is driving some regulation. On the flip side, technology innovation is creating applications consumers really want, and many are willing to let themselves be tracked to get access to them. This market momentum is driving even more innovation and adoption, so even more data is collected, causing privacy advocates to become even more fearful, driving even more regulation. And I think this dynamic is going to continue into the future.
Mike: What are we progressing toward—what’s your view of the future of identity?
Frank: I believe the future of identity is moving to mobile phones. We’re already seeing that with biometrics. Today there’s somewhere around $450 billion in payments being made using biometrics on mobile devices. That number is expected to reach $3 trillion in two years. It’s growing by leaps and bounds because people trust their phone; they know they have it in their possession. So I think iPhone and Android phones are the way digital identity is going to be enabled in the future.
Mike: When you think about the future, what scares you most and what excites you most?
Frank: Data breaches are my biggest concern. That’s one of those things that can really hurt a business in a big way—and it’s something I think every single company in the world has to worry about and do their best to protect against. That, and ransomware. The Colonial Pipeline as a recent example. They paid the ransom, but it’s been very damaging to the company as well as to a lot of consumers.
What excites me most is the billions and billions of dollars being pumped into fraud tech companies. We’re seeing an unprecedented level of investment activity. That’s really bad news for fraudsters because with all this money going in, and all these great people joining the fight, we’re going to make a big impact on fraud over the next 12 to 18 months.