Banks spent years treating payments modernization as a long-range technology project.
Corporate clients are done waiting. They want payments that move continuously, settle faster and sit directly inside their digital channels, and that pressure is now forcing banks to rebuild infrastructure they deferred upgrading for decades.
That was the central argument running through a PYMNTS panel moderated by CEO Karen Webster, featuring Barry Rodrigues, executive vice president of payments at Finastra; Peter Geronimo, executive vice president and head of PMI sales distribution at U.S. Bank; and AJ McCray, head of global payments product at Bank of America.
McCray put the timeline in perspective.
“This has been a decade-old conversation, I would say decades-long,” he said. “What we’re modernizing now was modern 10 or 20 or 30 years ago.”
Advertisement: Scroll to Continue
What changed is where payments sit. They are no longer back-office infrastructure. They are increasingly the product itself.
“You’ve got a new set of customers where payments in particular [are] much more embedded in what they’re offering to their clients,” McCray said. “And that really raises the bar for both how they’re interacting with their customers and what they’re looking for [their] bank to do.”
Putting the Client at the Center
The client relationship has been restructured. Banks no longer design payments infrastructure around their own systems and invite clients to connect. The direction has reversed.
“Historically, we’ve thought about how should the client connect to the bank,” Geronimo said. “Now it’s shifting to how should the bank fit into how the client operates.”
That shift reaches beyond treasury. Procurement departments, product organizations and customer experience groups now influence payments decisions because payments functionality affects how companies compete for customers. A U.S. Bank CFO survey found that 57% of respondents tied modernization to fraud and cybersecurity, 52% to better use of payments data and forecasting, and 42% to embedded payments.
The infrastructure underneath, though, has not kept pace with the client-facing work. Banks invested in digital channels during the pandemic but deferred the harder upgrades.
“The underlying platforms have not been modernized as much because it’s not so easy,” Rodrigues said. “Banks have got to keep the lights on with mission-critical systems at the same time as modernize.”
He also drew a line around what modernization cannot compromise.
“Payments are deterministic,” Rodrigues said. “They’re not probabilistic, so they have to work all the time, every time.”
Resilience, security, regulatory compliance and 24/7 availability are now baseline requirements, not differentiators, he said.
Questioning the Requirements
Webster used the discussion to introduce themes from “The Algorithm,” a book by former Tesla executive Jon McNeill that examines how Tesla and SpaceX built systems around outcomes rather than entrenched assumptions. The panelists said that framing fits the banking industry well.
Rodrigues said artificial intelligence and modern development tools now allow banks to rethink entire software lifecycles, not just accelerate existing ones.
“If you do it incrementally, bit by bit, what you’re doing is effectively just building a faster car as opposed to a spaceship,” he said.
McCray applied that logic directly to transaction processing. Legacy systems route payments through sequential hops before settlement. Banks are now redesigning those flows to run in parallel.
“What we’re looking at is how can you parallel process, how can you create multiple lanes to do the same activities,” he said.
The urgency is real. Clients no longer accept systems built around banking hours.
“Payments used to be nine-to-five, five days a week,” McCray said. “Now many of our clients expect payments on weekends and a 24/7 customer experience.”
Bank of America has rebuilt portions of its payments infrastructure from batch-oriented systems into API-based architecture supporting real-time payments across more than 50 countries.
Rodrigues pointed to modular architecture as another path forward, allowing banks to scale and modernize without replacing every component at once. U.S. Bank is running similar work through its SinglePoint platform, which consolidates dozens of capabilities into one environment for roughly 45,000 clients.
Geronimo said onboarding speed has become a top priority.
“Our biggest focus is on getting to a two-day onboarding process,” he said. “It all ties back to the client experience.”
AI Compresses the Timeline
AI ran through the conversation not as a future theme but as an operational tool already in use. Rodrigues described it as a force accelerator helping payments teams identify failed transactions, locate processing breakdowns and surface repair recommendations through natural-language interfaces, reducing operational workloads while freeing staff for higher-value work.
McCray said AI is reshaping how banks build products, not just run them.
“It’s completely changing the way we build,” he said. “It’s about how do you write requirements, how do you engage the operational partners, how do you interrogate what the legacy system used to do.”
Geronimo said U.S. Bank is applying AI to sales support, client analysis and operational efficiency, with governance controls in place around its use.
Looking ahead, the executives said they see broader adoption of instant payments, tokenized deposits and stablecoin infrastructure as the next phase. However, they added that modernization itself has no endpoint.
“We’ll still be talking about modernization,” Rodrigues said. “It’s just that the rate of change will be faster.”
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
