Paymentology CTO: AI Agents Must Earn the Right to Move Money

Paymentology CTO: AI Agents Must Earn the Right to Move Money

Payment infrastructure company Paymentology's Chief Technology Officer, Tim Joslyn, said in an interview with PYMNTS' "What's Next in Payments" series that as AI agents enter payment flows, the industry's core problem is no longer automation itself, but how to safely interpret and consume every decision made by machines.

Joslyn noted that existing payment risk-control systems are good at answering questions centered on human behavior—such as "Is this the cardholder?", "Is the account authorized?", and "Is the transaction fraudulent?"—while AI agents raise a different set of questions: even if a machine's identity is legitimate, does it have the authority to make this specific decision? He compared an agent's autonomy to an underwriting process, stressing that "autonomy must be earned through evidence," and that a model should not be assumed to be entitled to make decisions simply because it can.

On the question of liability, Joslyn argued that if an agent payment produces a transaction or decision that should not have happened, there is currently no clear framework for determining whether the responsible party is the programmer, the deploying enterprise, or the agent operator. He said technological evolution has outpaced the establishment of liability frameworks, and the industry needs to first define what each agent is permitted to do, rather than pursuing agents that can do everything.

Source: https://www.pymnts.com/news/artificial-intelligence/2026/paymentology-ai-agents-must-earn-the-right-to-move-money/