Worldpay is investing in more data-sharing techniques and artificial intelligence tools to fight fraud for its merchant clients, says Sunny Thakkar, head of global merchant fraud products at the payments processor.
It’s been a year since Worldpay was spun off from its former parent Fidelity National Information Services to become majority-owned by Chicago private equity firm GTCR. Now, Cincinnati-based Worldpay is tapping some of the up to $1.25 billion in growth capital that came with that transaction to better fight fraud and increase false credit declines.
Worldpay’s clients include mom-and-pop shops and mega-store retailers, with customers all over the world. The company processes about 2.3 trillion transactions annually. In this conversation, Thakkar talked about how Worldpay is working with retailers, card issuers and networks to share more data and use AI to improve merchants’ fraud-fighting capabilities.
Editor's note: This interview has been edited for clarity and brevity
PAYMENTS DIVE: What does fraud look like right now from your perspective?
SUNNY THAKKAR: There's a lot of vulnerability online because it's harder to capture that type of fraud when it's behind a screen. But that's where the volume is moving towards...fraud management is a double edged sword. So on one side, you want to stop and prevent the fraud, and on the other side, you need to prove as much good as possible. The problem is that the sword is much sharper on one side. What I mean is that fraud was $10 billion in 2023 for the U.S. [However] false declines — declining legitimate, good customers accidentally thinking they're fraud — is $150 billion problem.
What are some of the challenges in fighting fraud:
Refunds and returns abuse is extremely high and increasing, especially in the retail space. Since the [COVID-19] pandemic, people have gotten used to being able to buy multiple things and return them all online. They want free shipping, and free return shipping is the standard. It's very expensive for a business to operate that way, but it's also the experience that needs to be delivered.
How is Worldpay using AI to help its merchant clients fight fraud?
Machine learning is a component of AI that is able to make its own decisions based on the vast amounts of data that's provided to it. So that's an area that I think using as much data, good quality data, as possible is going to be where the companies are able to make the biggest impact on the fraud, which we're seeing continually evolve and grow, without impacting the good customers.
3D Secure is the widely-used fraud system, but what are its shortcomings with respect to liability shift?
Not all issuers use 3D Secure the way that it was designed to be used — that was left up to the issuer. So, you could have a transaction that comes in as a 3D Secure, you fully authenticate that as an issuer, but then you decline that in your fraud solution. So, it shows as a fully-authenticated transaction but the authorization is defined. So there are loopholes that issuers are finding to get around that, because the other part of the ecosystem that's causing that bad behavior on the issuer side is that there have been both merchants and other fraud providers that will take the riskiest of transactions and send those down 3DS and and try to get liability shift on that. So, there's basically an abuse of that 3D Secure technology happening because there's no mandate around it. So, it's not been a good roll out in the United States.
Is there a fix for that?
There's a new initiative, both from Visa and Mastercard, around what they call data-only, basically using the 3D Secure ecosystem not to introduce friction, or add a liability shift, but to share data. So being able to actually take data that's collected by the merchant, as part of the 3D Secure process, and package that data up and securely share that with the [card] issuer, and then the issuer has more information to make a decision, the liability doesn't shift in that and friction is not created either. It’s purely a data share.
What else is new on the fraud-fighting front?
We’ve started to create partnerships with issuers. We have one with Capital One right now. We're expanding to a few others this year. Once we make that fraud decision, we're actually taking that data that we're getting from that fraud decision, and just like I talked about in the 3DS flow, we have an API that we share with Capital One that securely packages that information so that by the time the authorization reaches the issuer, they can see the data that is linked to that transaction, and have more confidence in approving that transaction. So this data share ecosystem is both around stopping fraud, but allowing more transactions to get approved, to solve that big false decline problem as well.
