This episode was recorded at the Money 20/20 event in 2019. On this episode, PaymentsJournal’s editor-in-chief, Ryan McEndarfer, sat down with Aaron Lazor, co-founder and CEO of Finscend, and Moshe Teren, co-founder and CTO of Finscend.
Ryan Mac:
Welcome to the PaymentsJournal podcast. I’m your host Ryan Mac and in today’s episode, we’re going to be taking a look at credit card disputes during a conversation that I had with Aaron Lazor who is the co-founder and CEO of Finscend and Moshe Teren, co-founder and CTO of Finscend, during Money 2020. Now there’s certainly a lot to unpack during this episode, so without any further delay, let’s start the show.
Aaron and Moshe thank you so much for joining me on today’s episode. So to start off, what is Finscend’s key to technology?
Lazor:
Thanks, Ryan. Good to be here as well, Finscend’s technology, in a nutshell, is taking the arduous task that banks and credit card issuers have today of onboarding the information from a client for a credit card holder, just like any of us, and taking a process today, that could take banks up to three hours— these are industry statistics that we’ve seen—and condenses that down to just a few minutes. And on the consumer side, as a digital age that we’re in today, I expect to be able to make a digital payment by tap on pay within just a few seconds. I shouldn’t have to go through a whole, you know, long process with my credit card issuer to get resolution on any type of dispute that I would have.
So we use sophisticated technologies, we’re using NLP, natural language processing, to understand the specifics of the case and whether or not it’s authentic. And then we couple that with artificial intelligence to explore the merits of the case based on the guidelines of the card networks. The cardholder enters this information into a very straightforward mobile based or laptop based onboarding tool, just very simple information, questions and answers, uploading some documents that tell the application, our software, what is the merits of the case. The technology behind the scenes is using, as I said, NLP and artificial intelligence to look at the merits of the case and determine whether or not there is cause for a chargeback. And if so, it also produces a score—a score from one to 100—that gives the banks the information that they need to process this dispute.
Mac:
Alright, so now let’s get down to brass tacks here, right? So what is the problem in the marketplace that Finscend is solving for?
Teren:
Okay, thanks, Ryan. Thanks for the question. So the market right now for credit card transactions continues to grow year over year. Card-not-present transactions is expected to increase to $6 trillion by 2024. With card-not-present transactions increasing, the likelihood of a transaction resulting in a dispute is going to increase as well. There’s no longer the face-to-face interaction between the card holder and the person he or she is buying the product or service from. So this is leading to an increase in the number of disputes banks need to handle. In fact, the top 15 banks in the United States will spend over $3 billion a year just processing disputes. So we have to figure out a solution to make the process less painful for the bank. And in doing so we increase client satisfaction. And we keep the cards that the banks are issuing to the card holders at the front of their wallet, as opposed to in the back of the wallet, because frustrated clients today have a significant number of options when choosing their credit card. And they’re likely to choose that when something goes wrong, according to the bank that’s going to provide them the best service. So by expediting the way the complaints are handled, and minimizing fees, we feel very strongly that we’re solving a significant burden that the banks are dealing with today.
Ryan Mac:
Yes, certainly. I think that that’s very interesting here. And I think that you alluded to this a little bit here, but really, what is the return on investment for banks that use Finscend’s platform for credit card dispute resolution?
Lazor:
We’ve actually seen banks that use Google Docs in order to manage your disputes. So by using our full end to end enterprise-based solution, which we call the Bank Dispute Platform or the BDP, we anticipate the banks will be able to save up to 40, if not more, percent of their operating expenses. Not only that, there’s opportunity to create additional value for the clients, and clients that are happy with their credit cards tend to use that credit card more often. Think about it from a purpose of travel. If your credit card provides you travel protection, you’re more likely to use that credit card when traveling than another one that doesn’t have the same level of travel protection on it. So the investment here from the bank is multifold. Their clients are happier, your costs are reduced, you can repurpose your key employees to other roles within the organization. There’s the potential to minimize the need for third party processing. And everybody benefits because, quite honestly, we’re not looking to encourage disputes. We’re looking to ensure that disputes that are raised are valid, thereby eliminating invalid ones. And those that are valid, processing them more quickly so that the bank doesn’t have it sitting on its desk, and the client gets his or her money back in a credit card account as quickly as possible.
Teren:
I think also one of the key value adds to Finscend’s Bank Dispute Platform that we offer to the credit card issuers is that the enormous amount of time that they spend on trying to understand the merits of the case, we’ve spoken to banks and financial institutions, frankly, around the world, and the message is the same. The process for them is painful, it’s costly, it requires a lot of labor. Often they don’t have the customer service or dispute resolution teams onsite or employees of the bank, they offshore it or they use third party processors to onboard this information. And the effort of onboarding that information is not just timely, it costs a lot of money to the bank and the bottom line of the bank. It leads to different disparate results based on the idiosyncrasies of the customer service representative who’s looking at that particular dispute. We don’t want their employees to be judge and jury of dispute resolution. They want to be able to quickly see the information, judge the merits of it based on the guidelines of the card networks, and to come up very quickly with a resolution to that dispute.
Mac:
So I have to ask, because you had pointed out that you had seen some banks use Google Docs for their credit card disputes. That’s just not a marketing kind of thing for you to say, to kind of be like, oh, that’s provocative here. Like you have actually seen that, correct?
Lazor:
I hate to say it, but the answer is a big yes.
Mac:
Wow, that’s certainly very interesting. I never in a million years would have thought that, especially from a bank. Now, let’s take a look here at AI and machine learning, how is it that Finscend is using AI and machine learning in today’s platform?
Lazor:
We love this question, Ryan. So there’s a couple of points I want to get across here. First is how to actually build an effective AI. This is the first thing I’d like to talk about. The second thing I’d like to talk about is what companies are using their AI for today, especially in the payments ecosystem. So when you talk about AI, you talk about machine learning, you talk about millions and tens of millions and hundreds of millions of pieces of data that are flowing through a system in calculations, which are spitting out a score, or a value, or a recommendation based on how the creator of the AI thinks the data should look. So, for example, I want to try to find a solution to issue credit to somebody, how do I take that information and create a score which either allows me to feel comfortable to issue credit or not?
What separates Finscend from the companies that we’ve spoken with, and the AI solutions that we’ve looked into, is that more than just data manipulation, we take into consideration the client journey. What is the client feeling? What is his or her role in the transaction itself? By doing this, in building this into the formulas of the AI, we can see if the tendency of the client matches the dispute itself, thereby creating a more effective and valid dispute. Or, maybe it’s more of a random event. The bank then has the benefit of seeing the outcome and the recommendation created by our artificial intelligence predictive scoring model, and then can auto decision up to 80% of the chargebacks coming in, because they’re not just looking at data. Using an example, imagine every time you go on a holiday, you file a charge back on your hotel. So it would be prudent for the bank to know that this guy Ryan, every time he goes to his hotel, he seems to have a problem. This is not necessarily a hotel chain’s problem. Maybe it’s the way Ryan looks at the hotel and what he’s expecting from the hotel itself. So these pieces of information are included in the AI.
The second thing I like to bring to attention is the fact that in the payment space, companies are focused on this thing called friendly fraud. Friendly fraud, for those that don’t know, is just a scenario where two parties who don’t know each other encounter scenario where a purchaser claims he doesn’t know the merchant, right? So I buy something from merchant A, I received that product or that service, and I claim I never made a transaction. This is considered friendly fraud. I haven’t reported my card lost or stolen. So the power of AI today in this space is focused strictly on whether or not I’m making a legitimate fraud claim on a transaction. So companies that we speak with might ask questions like, is this the IP of your router at home? Do you have kids that have an iPad? Do you play Candy Crush? Is it possible one of your kids play Candy Crush? Maybe a transactions being disputed that is not recognized maybe by accident, but by asking additional questions and using artificial intelligence, they can kind of steer the client into remembering a relationship that the cardholder had with the merchant and therefore eliminate friendly fraud.
But as mentioned earlier, card-not-present transactions are increasing, which creates an additional exposure in the marketplace. So I can go ahead and purchase something online, never meet the actual seller of the product, he can be sitting in some other part of the world, could be a single man operation without a customer support department, whatever the case might be, and now we have to go in and interpret, based on the information provided by the cardholder, whether or not a transaction dispute is a valid dispute when the cardholder says, I made the transaction and I made this specific purchase, but for some reason, I’m disputing the transaction. It could be that I ordered a table from a carpenter in some other part of the country, and he has to ship it to me. And by the time he got to me there was dings and scratches. But I’ve never met the guy. So I have to hope that when I contact the factory, he says, I’ll send you another tabletop. So Finscend’s AI is not focused on friendly fraud, although that is a byproduct, and that is a space that’s, I guess, inundated with technology. It’s focused on this unique niche which is growing year over year of service and product related disputes. We get in there and by understanding client behavior, by understanding the merchant’s transaction record, the number of chargebacks that he’s had, whether the merchant has changed ownership so recently, we can create a confidence score, using artificial intelligence, that can point the bank in the best direction possible on the validity of a chargeback.
Teren:
There’s another great benefit to the AI in looking at the information that the consumer’s providing. The two great aspects that we’ve seen over thousands and thousands of cases that we’ve researched in becoming content matter experts in dispute resolution is that there’s a judging process that is given or put upon a dispute resolution customer service agent for a bank, to try to understand the myriad of rules and regulations and compliance. And all of these aspects of a case in a whole bunch of information that’s provided, some of it’s relevant, some of it’s not. AI uses science to break down, if you want to say to the brass tacks of the elements that are important to the dispute, and cuts through all the fluff.
The second aspect is that AI also can use the power of, I want to say, the uniqueness of that financial institution and to offer a recommendation on what should be the outcome of this result.
Mac:
No, I think that’s very interesting in the way that you’re implementing it. I mean, it makes perfect sense. If you understand the entire customer’s payment history and their behavior of it, it makes it somewhat easy to kind of say, okay, from a dispute standpoint, well, yes this individual buys coffee, usually between 10 and 10:30 every single morning, here’s the payment amount that they usually do. And why are they randomly now disputing every one of those transactions, that kind of seems to break their normal pattern of behavior here? So it makes it very easy to say, well, was this actually an illegitimate [transaction] or what is actually going on here? It paints a better picture, I think, overall than just kind of saying a he said, she said type of type a thing; there’s more data to be had there.
So all right, let’s say that I want to integrate and have Finscend come onto my bank’s platform here. What are the challenges that a bank is going to face with this and how easy is it for banks to implement Finscend?
Teren:
Yeah, that’s actually one of the important features of how we built the technology. The Bank Dispute Platform onboarding tool can be used as a standalone system or it could also be integrated into the heart or the backbone of a bank or financial institution; it’s really up to them. We built the system in a very modular way, that the onboarding tool can simply interact with any number of the bank’s existing CRM systems with a straightforward API. Or we’ve also developed an enterprise level CRM, a full featured CRM, that can that the bank can use to manage the disputes through the rest of the ecosystem process.
Lazor:
I just want to add that we don’t want our technology to be an obstacle for the bank to overcome when trying to improve a process which is a disaster right now. So I credit my partner Moshe on the way he thinks and the way he developed the system to allow very easy integration. We actually can integrate the mobile onboarding product within 72 hours through API; it is very easy. Client consent is on our side. So we feel very strongly about the product we have. And we don’t want technology, as much as everybody talks about integrating new technology, we don’t want that technology to be an obstacle for us to be able to get into a bank and start helping them immediately.
Mac:
Excellent. Now, for our last question here, it seems like numerous companies are working on tech solutions for their credit card industry. What gives Finscend kind of their advantage here in this marketplace?
Lazor:
Yeah, we spoke about this a little bit earlier. But I would say that AI can’t solve customer service problems if you don’t understand what cardholders actually desire. So most of the companies that are producing AI solutions or technology solutions for the payments ecosystem are not clients that actually understand the cardholder’s pain, the cardholder’s experience. So the advantage that we have is that we understand the cardholder’s experience. I think of a scenario where somebody, as a consumer, has a particular issue, and they create a product because of a pain point that they felt in that issue. We’ve seen a number of very large companies that have achieved success just based on this one singular event. We’ve actually walked through the credit card journey with thousands of different cardholders. We’ve done this representing cardholders with over 800 banks worldwide. So we have a very diverse sample set of what the cardholders are facing and how the banks are processing these disputes.
The other thing I mentioned is, again, when you think about what we’ll call “sexy” in the market today, we’re looking at technologies that help increase or drive revenue at the top line, to streamline certain aspects, to cross sell other products and services to cardholders. But if you don’t start to figure out where the leaks are in the buckets, with the amount of competitors out there with the credit cards and digital banks, in particular, out there, companies are going to spend a lot of money trying to recruit new clients. And then they’re going to walk out the door losing valuable dollars that were spent. There was a time when it was critical to companies, not only to recruit new clients, but to maintain the loyalty of those clients for the long term. And today, it’s becoming much harder because I can go on Google and in three seconds, figure out the best travel reward card or cash reward card or who’s got the best whatever. So the payment ecosystem right now is inundated with these companies that are either doing technologies to help increase revenue or to help merchants prevent friendly fraud. But where’s the client in the whole process? Where’s the bank, the issuing bank of that card, benefiting from this process? So we feel that we’re uniquely positioned to understand where the cardholder is and what he feels, as well as we’re providing a software solution to banks that help them to lower their operating expenses by 40 plus percent or more. And to keep the cards from the cardholders at the front of the card holder wallet, thereby increasing spending and thereby increasing revenue long term for the bank.
Teren:
The ecosystem is filled with companies that are looking for solutions on pre-credit authorization and on merchant side solutions. There’s very few technologies that we’ve seen that are focused on the issuing side and that interaction between the cardholder and the issuing institution. Our technology focuses there. And another key aspect is that we are trying with our technology to understand what the aspects of that dispute are and whether or not it should even become a chargeback. So we sit at the beginning of the payment ecosystem chain for those disputes.
Mac:
Well, Aaron and Moshe, thank you so much for taking the time today for speaking to me about Finscend, and I hope to have you both back on the podcast real soon.
Lazor:
Thank you, Ryan. We appreciate the opportunity.
Teren:
Thanks, and we’re looking forward to speaking to you again soon.