Payments Execs Say AI Makes the Payment Invisible and the Data Priceless

The Intelligence Layer Around the Transaction Becomes a New Prize

AI is reducing the cost of extracting meaning from enormous volumes of transaction data. That changes the competitive question from “Who processes the transaction?” to “Who owns enough context to interpret it?”

“The content of what someone is buying is far more relevant and valuable to that person and to those that are trying to provide services to them than the fact that they made a payment,” Jakob Harrison, senior vice president of payment networks at FIS, told PYMNTS.

An algorithm deciding what action should follow a payment needs context. An agent choosing between credit, debit, installments or rewards needs context. A financial institution trying to distinguish legitimate activity from fraud needs context. A supplier determining how aggressively to collect an overdue receivable needs context.

“The amount of data that we’re dealing with is quite profound,” Billtrust CEO Grant Halloran told PYMNTS. “There’s billions of events, behavioral signals and nonbehavioral signals as well, that now bring a data science lens to all of this. … Businesses are trying to generate the most cash possible from their receivables at the fastest rate and at the best economics.”

Payment infrastructure historically created value primarily by executing instructions. Today, the opportunity is to help determine what the instruction should be.

“Payments went from being just a line on a P&L [profit and loss statement] to the thing that can actually drive product roadmaps [and] customer retention,” Ron Griswold, director of strategic alliances at WEX, told PYMNTS.

AI Makes the Software Cheaper and the Moat More Important

Still, one uncomfortable consequence of artificial intelligence is that it looks to be improving both incumbents and challengers simultaneously.

Companies “are changing so much that a new competitor may come from a completely different kind of organization,” João Del Valle, co-founder and CEO of EBANX, told PYMNTS, adding that AI-native firms can operate with different cost structures, development cycles and organizational requirements than conventional payments companies.

“Five years from now, we won’t have even an idea of what a payments platform will look like, given the amount of innovation that’s going to happen,” Del Valle said.

Capabilities that previously required years of engineering, large development teams and significant capital can become faster and cheaper to reproduce. A competitor doesn’t necessarily need to replicate an incumbent’s organization if AI allows it to replicate enough of the incumbent’s product.

“AI improving productivity, to me, that’s table stakes,” Nandan Sheth, CEO of Splitit, told PYMNTS. “The real threat is someone using AI to remove entire layers of friction, cost and intermediaries within my segment. … The question I ask is, what would someone build today if they weren’t constrained by our existing architecture, organization and assumptions?”

The exercise exposes the other side of AI commoditization. If software becomes easier to reproduce, assets that a model can’t easily generate become more valuable. AI may therefore create a strange reversal in technology strategy. The more ubiquitous sophisticated software becomes, the more competitive advantage migrates toward assets outside the software itself.

The Next Checkout May Not Have a Checkout Page

The most visible manifestation of that shift may eventually be an invisible payment. An AI agent that understands a consumer’s financial preferences could potentially determine how to fund a transaction before the consumer encounters a checkout screen.

“The concept of agentic commerce, of agents doing the shopping for you, is like nothing we’ve seen before,” Maran Nalluswami, executive vice president and chief strategy and business development officer at Synchrony, told PYMNTS. “You almost have to place bets along the entire experience right now because the experience has not been won by anyone yet.”

“I don’t think the inflection point’s moving so fast that you can’t adapt to it,” he said, “but it’s important that you stay ahead.” If consumers increasingly begin shopping with an AI assistant rather than a retailer, search engine or marketplace, then the entity that mediates discovery gains influence over the presented merchant, product, financing offer and payment credential. Payments could consequently become more invisible to consumers while becoming more strategically important to the platforms coordinating commerce.

“A technology cycle is something you adopt,” i2c Chief Client Officer Jason Goldberg told PYMNTS. “You buy the module, you check a box, you move on. An inflection point is dramatically different. It changes the architecture of how you compete in the market.”

And in agentic commerce, the payment does not disappear economically. It disappears experientially.

“Payments are being absorbed by commerce,” Splitit’s Sheth said. “They’re becoming invisible within commerce.”

Governed Speed Is Becoming the Real Moat

If intelligence is becoming more important and AI is accelerating development, another scarce capability emerges: the ability to move quickly without losing control. “You can’t do a blanket decision that we’re going to do real-time payments for a huge block of our outgoing payments,” WEX’s Griswold said. “The faster those payments go from one bank to another, one company to another, the less due diligence can be done in between.”

Payments companies and financial services firms cannot simply maximize velocity. They simultaneously have to preserve reliability, security and trust.

“If you move into regulated moments without the appropriate guardrails, the compliance controls, the QA, and the auditability, you’ve traded a competitive problem for a regulatory or trust problem,” i2c’s Goldberg said. “Trust is earned slowly, and it’s lost quickly.”

The value as a result shifts from possessing a faster rail toward knowing when to use it. The same idea applies more broadly. Systems that can understand an event while it’s happening have an advantage over systems that can only explain it afterward.

Chris Phillips, director of Financial Crime Industry Engagement at Flagright, said institutions need intentional governance and an audit trail running “from data lineage sources up through decisioning.”

“You can’t just tack on [things like] crypto on top of a typical rules-based system, and say, ‘Hey, hopefully it works,’” Phillips said.

Real-time settlement that accelerates an existing payment improves the product. Real-time information that allows companies to make financial decisions before a bank statement arrives changes the operating model.

“You can’t have a knee-jerk reaction to anything,” Nalluswami said, noting that what emerges instead is a management discipline built around separating signal from noise.

The biggest risk facing payments companies isn’t necessarily missing a technology. It is misunderstanding which assumptions the technology has already invalidated.