In an era of rapidly evolving technology, few industries are as poised for transformation as banking and payments. This is a transformation already underway with technologies such as generative artificial intelligence (GenAI)..
“Governance, cost and The utility is formed by the three points of the triangle that we are trying to balance (GenAI). If we get it right, it opens up possibilities for transformation in our industry. ” Mark SantoCTO payment of taxessaid in a conversation on the PYMNTS series “Next Steps in Payments: A Note to GenAI Companies.”
As banks and payment companies consider this technology, the focus is shifting to practical applications of AI that can increase efficiency and improve the customer experience. and Address fraud.
With AI tools embedded in almost every software service, businesses must also prioritize guardrails to ensure data privacy and security.
While the potential of AI is undeniable, Sundt highlighted significant concerns regarding safety and cost. “Governance is our role beginning “It’s a priority,” he said. “Every service we use has some kind of AI component built into it. Governance is about making sure these tools don’t inadvertently spread sensitive data.”
Challenges of AI payment applications
Equally pressing are cost considerations for the use of AI. Large-scale language models (LLMs) are powerful, but their operational costs can be prohibitive, and their usefulness does not always justify the investment.
But as the industry moves away from monolithic models that seek to solve every problem, smaller, specialized “agent models” are gaining traction. These models are designed to address specific problems efficiently and cost-effectively, changing the way AI is deployed.
“We are moving from the kitchen sink approach of large-scale models to centralized, organized agents that solve specific problems effectively and at low cost,” Suunto said.
This trend is especially evident in customer service, where AI is being used to accurately diagnose and resolve issues, he added. Rather than building large models that try to answer every possible question, companies are developing smaller models that focus on solving specific problems.
“Customers don’t want encyclopedic answers. They want to know how to configure their terminals to optimize interchange fees and response times. That’s where small, specialized models shine. ” said Suunto.
“This pivot toward a dynamic, compact model is similar to the evolution of computing itself,” he adds, as it parallels a broader shift in technology from mainframe computing to distributed cloud applications. He pointed out that
Fraud detection and risk management using AI
The role of AI in risk management and fraud detection is another area with immense potential.
“The biggest red flags we encounter are merchants with newly established banking relationships or websites. These temporal attributes often indicate fraudulent intent,” he said. I did. Stax leverages AI to enhance decision-making by evaluating factors such as account age and in-store presence. and transactional behavior.
Suunto also described suspicious patterns in transaction fraud, such as large transactions followed by bulk reversals or refunds issued to different credit cards. “These scenarios require robust AI systems to detect and mitigate fraud at scale,” he said.
Looking ahead, Sundt believes the next frontier for AI will be in inference.
“Current models are great at summarizing and classifying, but struggle with inference. Humans always make decisions based on incomplete information. Teaching AI to do the same is difficult. “It will be a breakthrough,” he said.
This capability powers everything from fraud detection to customer service, allowing AI to better navigate complex and ambiguous scenarios.
Still, Sundt emphasized that AI providers need to adjust their rapid development cycles to the slow, compliance-driven nature of the banking industry.
“The financial sector does not move at the speed of technology. It is very important for AI companies to understand We consider the unique challenges of our industry and design solutions accordingly,” he said.
