Picture a future where paying for everyday needs could happen without a tap or a click, as artificial intelligence starts to move from recommending products to handling transactions on your behalf. DBS Bank and Visa are testing this possibility, piloting a new system called Visa Intelligent Commerce, where digital agents can order and pay for items using customer payment credentials, under the bank’s oversight. Food and beverage purchases have already been completed automatically using DBS and POSB cards, pointing towards a new phase in digital commerce. The trial stands as one of the first public examples of banks actively exploring AI-driven payments in live market conditions, offering a look at how human involvement in payments might gradually lessen over time.
Announcements surrounding DBS’s AI efforts have frequently centered on operational improvements and customer-facing digital interfaces in the past. However, most previous news reported automation for internal use, such as streamlining support or optimizing credit scoring, rather than direct payment execution. Earlier commentary often focused on fraud prevention and chatbot support, without actual purchase completion. The move to allow AI to finalize transactions shifts the conversation from support tools to practical, everyday use, directly affecting customer experience at the point of sale.
How does agent-driven commerce work?
AI-driven commerce, as tested by DBS and Visa, puts digital agents in charge of the entire transaction process after receiving user directions. The system allows automated agents to search, select, and buy products, but all operations are subject to strict controls. Payment data is tokenized and passes through validation steps, ensuring the customer’s intent and authorization remain at the center. According to Visa, this approach is meant to safeguard against unauthorized activity, while still giving users a streamlined purchasing experience.
What types of purchases are being tested?
Initial implementations focus on routine expenditures like grocery shopping, food delivery, and subscription renewals. Customers can set parameters such as preferred retailers, budget ranges, and order frequency before the AI executes transactions on their behalf. DBS and Visa intend to widen the scope of the pilot to cover more areas, including travel bookings and broader online commerce. The companies aim to identify where customers feel most comfortable letting software handle orders.
What are the risks and benefits for banks and consumers?
Agent-based systems could let banks maintain their role as digital gatekeepers, overseeing AI-driven spending and approving transactions based on customer preferences. However, there are questions about who is responsible if an agent purchases something the user did not intend. “Our primary goal is to ensure that the customer remains in control, even as AI automates certain tasks,” said a DBS spokesperson. Visa emphasized the importance of maintaining safeguards, stating,
“We’ve built security and consent protections at every step to build user trust in AI-initiated transactions.”
These safeguards, centered within the bank’s infrastructure, may encourage cautious acceptance of AI-powered payments.
AI’s direct participation in financial transactions marks a meaningful shift for banks, moving beyond marketing suggestions or chatbots to acting on customers’ behalf. The success of such systems will likely come down to user confidence, the reliability of approval flows, and banks’ clarity around setting spending boundaries for agents.
“By piloting Visa Intelligent Commerce, we’re exploring tangible uses for AI in everyday banking beyond just simplifying tasks,”
the DBS representative added. Early use cases, as planned, are likely to stick to low-risk, frequent purchases until trust develops for more complex orders.
Consumers seeking convenience may find value in delegating routine transactions to reliable software, but bank-controlled safeguards will be crucial to minimizing errors and unauthorized spending. The entry of AI into commerce introduces new questions on data security, liability, and consent management, but could eventually lead to widespread use where AI agents handle significant portions of retail transactions. As with any digital advancement, user education, clear consent protocols, and transparent resolution processes will be key for adoption. Businesses considering these tools should evaluate risk profiles, customer readiness, and operational capacity before implementation. Those keen on experimenting with AI payments can monitor pilots like DBS and Visa’s, which may soon influence how institutions around the world approach everyday digital spending.
