How Could Generative AI Change Payments Processing?

While generative artificial intelligence (AI) makes headlines, the technology has for years been driving efficiencies.

Particularly within the payments space.

Already, automated machine learning (ML) and predictive AI solutions are helping firms streamline formerly manual processes within areas like accounts payable (AP) and accounts receivable (AR), cash flow forecasting, credit scoring, fraud prevention and compliance.

Areas like these will be the “easiest and first” avenues where applications of generative AI can score an immediate impact, Tom Randklev, global head of product at payment orchestration platform Cellpoint Digital, told PYMNTS.

“Generative AI can essentially retrain the old AI and machine learning models,” Randklev explained, adding that he sees a “really interesting trajectory [for AI] when it comes to payments.”

That’s because given the exponential growth of eCommerce and digital, embedded payments, today’s firms need to not just protect themselves against a rising tide of modern fraud, but also find new ways to acquire and retain customers within an increasingly digitized commerce landscape that is constantly evolving.

“Consumer convenience plays to the very front of this one, where embedded payments have already made the experience incredibly smooth — it’s a one- or two-click experience to get from product selection to a payment, and I think generative AI will continue to accelerate that,” Randklev said.

Read moreGenerative vs Predictive AI’s Role Across the Future of Payments

Powerful Enabler of Convenience

Beyond just helping firms uncover new opportunities based on an improved ability to process and generate insights from vast troves of data, generative AI applied to business processes can also help boost revenues through increased personalization of services.

“[Generative AI will impact] everything from identity verification to nontraditional credit scoring, this is another evolution step in contextual commerce allowing merchants to meet their customers where and how they want to do business,” Randklev said.

By leveraging different types of AI, companies can unlock insights into consumer behavior.

Based on predictive modeling around purchase patterns and more, businesses will be able to provide forms of payment that are more meaningful and increase conversion, Randklev added.

“When you have a vast ecosystem of capabilities and connectivity, it allows for the most relevant processing partnerships that ensure transactions go through securely and go through with the greatest chance of achieving success,” he said.

Still, “it’s very early days,” he said, noting that the massive number of tools leveraging generative AI that are coming out each week “would’ve previously required a two- or three-year window” to start seeing the benefit of, but can now generate efficiencies almost immediately.

Tool of Transformational Potential

AI solutions, broadly speaking, can improve any process — but the tech’s capabilities really shine when applied to complex processes with large data volumes.

The history of AI development has been driven by the all-important fact that large data sets gathered for one purpose may yield potential new kinds of commercial knowledge because of computation and practical analysis.

“For years we’ve used some sort of AI and ML to drive capabilities across financial technology and payments — but a lot of that was written for purpose, and while predictive and a good way to digest data, it still has its limitations when compared to these new large language model-driven algorithms,” Randklev said.

From a payment standpoint, he added, the technology can accelerate optimization of routing to ensure there are greater approvals, drive cost control across an ecosystem, and transform fraud prevention and risk-related controls.

Still, Randklev said the “biggest influence” of AI will be its ability to optimize the customer experience in a way that increases conversion by creating efficiencies across the payment occasion “even from a cross-border standpoint.”

Given all of these immediately realizable efficiencies, the question is just what role, if any, humans will play.

But not to worry, Randklev said — there will always be the need for a “human in the loop.”

“It’s easy to put on your tinfoil hat … but ultimately there’s always a need to trust but verify, and to keep the human in the loop. While it is still early days, the accuracy of generative AI still leaves quite a bit to be desired,” he said, adding that many security and data privacy elements inherently need a human element to act as the control mechanism.

“If anything, [AI] just upgrades us — it makes us more efficient, and we cut the dead weight and hand it off to the robots, if you will,” Randklev said.