AI Becomes the Human-Machine Bridge That Makes Payments Safe and Smart

As the world buzzes about ChatGPT, the payments and financial sectors are already sizing up advanced new artificial intelligence (AI) tools relating to how they can move from parsing vast amounts of data instantly to detecting fraud to having human-like service interactions.

Speaking with PYMNTS for the Nuvei PYMNTS Merchant Series on the application of AI innovation in the payments and financial services space, PayU Global Payments Chief Product Officer Daniel Cohen said he sees a world of possibilities around AI, from the back office to consumer-facing uses.

Nodding to the growing use of AI-powered tools in areas like virtual try-on and in making informed product suggestions on the consumer retail front, Cohen said that as it pertains to finance and payments, the utility of AI starts with its abilities to spot things humans often don’t.

“It all comes down to how we use data,” he said. “As humans, we’re very limited at using data, at looking stuff up, at running complex statistics and trying to find maybe the right loan, the right investment. With AI and the simplicity that it introduces … the amount of data that this AI entity can collect, can process, and then feed information back to you is endless.”

Along these lines, PayU has said that nearly 70% of FinTechs will be significantly impacted by AI in the coming decade, and Cohen said he sees that activity concentrated in some core use cases.

“From a machine learning perspective, we’ve been using data for many years to help guide different decisions that happen in the back-end operation, [especially] around the orchestration layers that we have,” he said. “What do we use to process the transaction? Which acquirer do we go to? Which processor do we engage with? It’s helping with decisions that make the transaction complete successfully faster, and as cheaply as possible for our customers.”

See also: PayU Appoints Keren Ben Zvi as Head of Data

Leapfrogging Oceans of Data

“We’re constantly in this race in the online payments world to make things easier, more accessible, more transparent, and for the user, more trustable, more accessible,” Cohen said on the topic of FinTechs and AI. “AI leapfrogs a lot of the challenges that we face in the industry in making data more accessible and more understandable.”

That can mean anything from an AI using its data-crunching capabilities to suggest the right investments and loans to detecting fraud, all while reassuring humans that the fit is correct.

“When it comes to investment decisions that you’re trying to make and loan decisions, AI can help you quickly understand, in language that makes sense to you, what is at hand, what are the implications of your decision, and then help you make that right decision,” he said. “When you ask about the most significant impact, I think it’s that.”

Companies are already using AI to improve back-office operations and ensure the best outcomes, especially where reams of sensitive financial data are involved.

“When we think about AI, kind of the building blocks of [it] are … primarily data, how we process data, and buzzwords such as big data and analytics and advanced analytics and machine learning,” Cohen said. “All that has existed for many years, and specifically with fraud, the anti-fraud industry has been doing a lot with data for many years.”

AI takes that to the next level, which is a positive boon to payment processors like PayU that operate in different markets and geographies and must provide consistent levels of service.

With fraud detection among the prime AI use cases, Cohen added that “our automated fraud detection mechanisms are pretty far advanced, where AI can now help beyond the detection to the triage, do the more mundane tasks that fraud analysts do, and help the fraud analyst focus on the bigger issues at hand, acting as an operator in the background to finish tasks.”

Read also: Payments Localization, BNPL and Cross-Border Drive PayU’s 3-Year Roadmap

Triaging Service Issues With AI

While PayU has been using machine learning and algorithms for many years, feeding on very large datasets and applying advanced analytics, AI’s customer-facing uses are also critical.

“From a back-office perspective, it’s about how we help customers and the support workload that’s coming in,” Cohen said. “Again, AI can be very helpful in triaging in a very human way. It’s not [a customer] waiting on the phone, press one, now press five, and it’s 30 minutes later and you still haven’t spoken to anyone. With AI, you’re immediately talking to a ‘human,’ and they’re answering your questions very directly and relevantly.”

“If it escalates, they can come back to a human,” he added. “That’s kind of where things are headed for us.”

While saying there will always “be human eyes on monetary operations,” Cohen got to the core of what AI does in finance, payments and retail to assist human-machine collaboration.

“It’s easing that human-machine interaction,” he said. “It’s a bridge, and it allows us to be very human in that engagement, to use our human skills to interact with a machine. That’s something that technology is always trying to make easier. It’s trying to make our interaction with machines and data as easy as possible. AI is a significant bridge to help us interact with machines in a more accessible and human fashion.”