How Merchant Fraud Detection helps Viva Wallet hypergrowth

July 14, 2022

Fraudio recently conducted a webinar, where experts discuss how can merchant fraud detection systems help merchant acquirers achieve hyper-growth, without the need to scale their fraud teams or sacrifice security.

This panel is moderated by Diederik Klopper, consultant at PaymentGenes with the participation of Giorgos Gkionis, Viva Wallet's Chief Software Architect Risk, AML & Anti-Fraud, and João Moura, Fraudio's CEO. They discuss the secret behind Viva Wallet’s evolution into a unicorn, and how Fraudio helped them achieve hypergrowth.


Webinar Transcript

Diederik Klopper (00:23)
Hi, everybody, and welcome to today's Webinar. Today we'll discuss the secret to hypergrowth for merchant acquirers. We have an interesting discussion between Viva Wallet and Fraudio. As you know, Viva Wallet is a leading European cloud-based neo bank. They offer unifying yet localized payment solutions and embedded banking services in over 24 markets across the continent. Representing Viva Wallet, we have George Gkionis. He has over 22 years of experience both as a software engineer and a software architect. And on the other side, we have Fraudio, which connects merchants, payment service providers, merchant acquirers, card issuers and other payments in the payment value-chain to a powerful and centralized AI smart brain, and that prevents, detects, and fights fraud in real-time. George, can you hear me?

Giorgos Gkionis (01:13)
Yes, of course.

Diederik Klopper (01:14)
Yes. All right, perfect. Representing Fraudio, we have João, a CEO and CTO that has over 14 years of experience working in AI. And he entered the payment industry in 2017 and founded Fraudio in 2019. So, first of all, a warm welcome. Thanks for joining. To start off, George, can tell us a bit more about Viva Wallet and what it is that you do within the organization.

Giorgos Gkionis (01:44)
Viva Wallet is a neo bank. As you mentioned, it started primarily as a payment service provider. It's a service that is still served by us. We are currently operating in 24 countries, offering acquiring services, issuing services, and pretty much everything that the company needs in order to perform any card transactions and payments. And we're a principal member of Visa and Mastercard, meaning that there is no in-between us and the schemes themselves, Visa and Mastercard. I'm the chief architect of our in-house anti-fraud solution. Pretty much every software framework that we have is in-house developed by us. We don't rely on any third party framework in order to perform all necessary operations concerning the payment service provision. And somewhere between there, is taking place our calls to Fraudio and their callbacks to us, offering us a pretty much reliable service that we will analyze in the following minutes.

Diederik Klopper (02:51)
Yeah, perfect. If you can also just tell us a bit more about why you started Fraudio and where you are right now.

João Moura (02:58)
Yeah, thanks for having me here Diederick and Giorgos. So a bit about Fraudio. We started this journey about three years ago, and we started it a little bit out of frustration. So we were working for a dutch acquiring bank and global payment service provider, and I was the head of Data Science there, and I was responsible for building products that had something to do with data and data science and AI. And I had to develop a fraud solution. We were working with a couple of fraud solutions that exist in the market, but they weren't suitable. So I had to basically develop our own with my team. And it turned out to be pretty good. And we're realize that we had a different angle or that we could have a different angle and really productize our solution instead of taking the approach that everyone has in the industry, which is very service-based, and fast forward three years and here we are.

Diederik Klopper (04:19)
Indeed, fast forward and time flies. George, today we'll discuss the secret to hypergrowth for merchant acquirers and then specifically the balance between customer experience and security. Take us back a few years before you started collaborating with Fraudio. Can you elaborate a bit on what the processes looked like at Viva Wallet and where the main bottlenecks were that you were trying to solve with the collaboration with Fraudio?

Giorgos Gkionis (04:57)
First of all, customer onboarding and business onboarding, in particular, is quite an issue, still. Let's rewind a few years before. The customer onboarding experience for us was pretty much straightforward. We were asking for very little data compared to now, from our customers. And once they were onboarded, they could perform transactions at pretty much everything that we had available. But then on a few years later, the Central European Bank and the bank of Greece told us, okay, you need a full KYB solution before a merchant can perform any transation whatsoever. It became quite complicated, especially in terms of security, and it was a problem. And we needed to find the balance between rapid customer onboarding and secure screening of our customers, as early as possible. So right now, we need as little data as possible from the customer, keeping in mind what the Central Bank of Europe and the Central Bank of Greece need in order for us to be compliant with them. And from the moment a customer is onboarded and the minimum amount of documents is requested from him, every information concerning his behavior is sent over to Fraudio. Fraudio is able to perform a complete customer behavioral analysis from day one and can inform us very promptly if anything goes wrong, if the customer is trying to do anything fishy, if there is an outlier anywhere in the customer journey. So in very few words, that's how we operate right now.

Diederik Klopper (07:03)
All right, João, you receive the data from Viva Wallet, and then you analyze it to see, where are the outliers, where are the inconsistencies? How do you analyze that? What are the parameters that you're looking at and how is that set up?

João Moura (07:18)
Yeah, so regarding specifically Merchant Initiated Fraud detection, we consume a lot of different money flows from Viva Wallet. So we consume not only everything that has to do with card payments, but also bank-to-bank transactions, wallet-to-wallet transactions, even cash withdrawals. So all the money flows that are seen by Viva, we consume them, and we also consume the information that is gathered at onboarding time. So during KYC / KYB, we receive all of that in real-time, all of those transactions in real-time, and the KYC / KYB information in batch. And then we look for each merchant, the unit of analysis for this product is the merchant. We look at that merchant and the very first step is really to put it together or to group it with similar merchants: so same MCC codes, same region, same type of profile, risk profile. But this is done automatically and by our AI. So it's not like we simply have a few rules -

Diederik Klopper (08:44)
- manual input that on the merchants together.

João Moura (08:47)
It's a lot more developed than that, a lot more complex than that. And so we end up really having groups of merchants that should behave similarly, and then sometimes they don't. Sometimes there are deviations from a merchant's processing patterns with respect to the merchants peer group. And that's really at the core of the analysis that we do. And then we do all types of anomaly detection and look for anomalous behavior.

Diederik Klopper (09:26)
Yeah. And I can imagine that works on two fronts. On one side is it greatly increases the batch size of what is a normal transaction so that you can identify the outliers. But of course, if all those outliers concentrate on one merchant, you're not enough just analyzing the data of that merchant, so also, therefore, you need a bigger group set of data.

João Moura (09:51)
Correct. And it actually goes beyond just transaction per transaction analysis. So the patterns that we look for can be really complex. So things like the typical velocity checks, those typical patterns where an entity has, say, 100 cards that they are trying to exploit, they typically will start by probing and seeing from those cards which ones are valid, which ones are selective, or which of those credentials that they have are correct. Let's put it that way. And then they will try to do, for instance, small transactions. Then some will file, some others will go through, and then they will set aside the ones that went through. And then they'll try to take as much money out of it as possible. These are very simple patterns that we track. But it already alludes to the fact that we don't look at transactions individually. So it's not like we see one card being used, one transaction here, and it looks fraudulent, and so we stop it. No, it's a lot more complex than that. And it's cross merchants also, and even cross acquirers. So we group merchants that we see that are being processed by Viva wallet, together with merchants from other acquirers.

João Moura (11:36)
And then sometimes we see the same set of cards being used across multiple acquirers and specifically for a few or in a few merchants. And then we know that those cards have been exposed, have been breached together. They are, let's say, a unit or a set of compromised cards and being used in a set of malicious merchants. And then the combination is very powerful and it really allows us to detect very precisely in a way that's not trivial, that can hardly be done by humans with spreadsheets.

Measure results yourself !

How about trying our solution  and experiencing the next generation for yourself?