Thredd CEO Jim McCarthy on Credit, Compliance and Agentic AI

Jim McCarthy has been in payments for 25 years, which means he’s seen processors come and go, watched prepaid transform fintech and knows exactly why credit is hard. Speaking at Money20/20 in Las Vegas, the Thredd CEO explained why his company waited this long to enter credit processing in the United States and why getting it wrong would hurt clients more than the processor itself.

Thredd built its business on prepaid, which Jim describes as the foundation for most fintech innovation over the past two decades. “A lot of the companies of our era came on the back of the prepaid revolution of the early 2000s and prepaid is a wonderful invention,” he says. “I call it the Swiss Army knife of payments. Most of the fintechs you see here at Money20/20, most of the innovation of the last 20 plus years, was built on the back of that innovation.”

But prepaid has limits and Thredd’s clients kept running into them. The company has now launched both debit and credit ledgers in the American market. “To service the needs of clients with increasingly complex embedded finance and commercial banks in a compliant way, we needed both,” Jim says.

The question is why now and how Thredd plans to sit between the slow incumbents and the fast-but-fragile startups without becoming either one.

Why the credit processing market has a gap

The credit processing space has what Jim describes as a chasm in the middle. On one side sit FIS and Fiserv, what he calls the three-legged stool of incumbents. On the other side are startups that move fast and break things, often at their clients’ expense. “Traditionally, the big folks in the space, they do what they do very well.

"They’re very good at credit, but they’re very bad at innovation. There’s a long tail of underserved issuers in the United States and around the world that are just dying to go faster.”

The other side presents different problems. Newer entrants move faster but lack credit expertise, and when they get things wrong, their clients pay the price. “On the other end, the newer entrants to the space, they're not very good at credit. It's complex, it’s hard to deliver.”

That’s where Jim’s thinking diverges from typical startup logic. When a credit processor fails, the damage doesn’t stay contained. “The scars and scar tissue occur because the clients learn that the processor’s not very good at this and they’re taking the wounds with their customers and then eventually with the regulators. We don’t want to do that.”

Thredd’s approach with LoanPro avoids learning on clients

Thredd partnered with LoanPro for its credit launch rather than building loan origination and servicing internally. The decision reflects Jim’s view that credit requires expertise you can't develop quickly and mistakes cost clients too much.

“We don’t want to learn on the job with credit, so we’re partnering with a great partner. What we’re bringing is best-in-breed card capabilities. That’s what we do really well. We’re going to leverage the loan origination and loan servicing capabilities of LoanPro and we’re bringing our credit ledger.”

The split makes sense. Thredd handles card processing, LoanPro manages loan infrastructure and clients get compliance from day one. Jim's been in the business long enough to know that service delivery matters more than speed when it comes to credit. “We want to be bank grade out of the gate because I’ve been in this business for 25 years. Credit’s tough. We know service delivery is everything – it’s one of our key selling principles.”

The platform model Thredd uses is horizontal, meaning clients bring problems and Thredd offers APIs to solve them. Credit expands that model into financing scenarios Thredd couldn't touch before, which changes how the company can respond to client requirements. “It truly opens up the spectrum of things we can do,” Jim says.

What embedded finance clients actually need credit for

Thredd’s fintech clients were already asking for credit before the company could provide it, which meant watching potential business walk away to competitors. “Our clients have been constrained because we couldn’t offer that. As you think about the world of embedded finance – enterprises, B2B clients – they want to be able to offer credit to their clients so they can effectively expand the open-to-buy for their customers.”

The use cases start simple: Buy now, pay later, earned wage access and overdraft protection, for example. These work for prepaid clients and neobanks extending into credit.

Jim mentions Pockit in the United Kingdom, which serves underserved customers and wants to offer short-term loans based on deposit patterns. 

Then there are commercial banks, the ones that can't get what they need from payment processors like FIS or Fiserv. These institutions want to move faster but can't on their current platforms. “They want to get on the front foot with innovation to compete against the tech companies that have come into the space. They can’t do it on those platforms.”

Banks need more than innovation, though. They need compliance baked in from the start, which is where many processors fail. “You’ve got to have compliance and you’ve got to take the risk out of the equation. You can’t just be about innovation – you’ve got to have that focus on doing it in a way that’s bank grade.”

How agentic commerce changes what processors need to do

AI now influences every aspect of how Thredd approaches its business. “You can’t be sitting in 2025 and not have a discussion about every aspect of our business without thinking about AI,” Jim says.

The immediate applications in back office operations are straightforward enough, with chargebacks, disputes and operational processes all benefiting from AI implementation. But Jim’s focus extends to the front end of payments, particularly around digital transactions and tokenisation, where infrastructure changes become more complex.

He draws a parallel to 2014, when he worked with now-Mastercard CTO Ed McLaughlin on the launch of Apple Pay. The consumer-facing product received significant attention, but the underlying tokenisation infrastructure represented the more substantial shift. “At the time, everyone was more focused on Apple and Apple Pay, but I think Ed would agree with me that tokens was the main message,” Jim says.

Visa and Mastercard have issued more than 40 billion tokens since then. Agentic commerce represents the next phase of this infrastructure evolution, where AI agents conduct transactions on behalf of consumers. “Agentic is the next wave of it, and it’s just unleashing a whole new wave of use cases, all of which drive commerce into the future in a more safe and sound manner,” he says.

The technical requirements involve new protocols for identity binding using passkeys and agent tokens. Thredd is working with Visa and Mastercard on implementing these protocols as they develop. The challenge centres on enabling commerce when the transaction originates from an AI agent rather than a direct consumer action.

“As we think about the future from an AI perspective, not only is there expense savings and efficiency savings in terms of the core processing aspects, we’re already thinking about how we work with Visa and Mastercard as they introduce the new protocols that drive new ways to bind identity using passkeys and agent tokens to make commerce more easy and seamless. We work with our partners and our clients to allow them to come up with great new use cases that leverage all these new innovations and protocols.”