Artificial intelligence is advancing beyond offering advice to directly executing actions, a trend highlighted by DBS Bank’s new pilot program. In partnership with Visa, DBS is trialing Visa Intelligent Commerce, a technology framework that enables AI agents to autonomously complete purchases for customers, signaling a transformative approach to everyday payments.
AI Agents Taking Control of Transactions
The pilot system empowers digital assistants to search for products, make selections, and finalize purchases using payment credentials managed by DBS. This marks a move from AI as a mere advisory tool to an active participant in financial transactions. Early real-world applications of the system include processing payments for food and beverage orders using DBS or POSB cards.
Ensuring Security and Customer Control
Visa’s framework emphasizes bank-centric oversight by tokenizing payment information and routing transactions through issuer-controlled approval processes. These measures confirm user identity and enforce spending limits, ensuring that AI agents operate strictly within the parameters set by customers and the bank. This approach balances automation with the necessary human oversight to maintain trust and security.
Focus on Routine Purchases and Gradual Expansion
Initial use cases for the AI-driven payment system target routine transactions such as grocery orders, subscription renewals, travel bookings, and household restocking. AI agents adhere to user-defined rules including budget constraints and brand preferences. DBS and Visa anticipate expanding the pilot to cover broader e-commerce and travel-related purchases as the technology matures.
Implications for the Financial Industry
The introduction of AI agents capable of making payments presents both opportunities and challenges for financial institutions. Banks that adopt agent-based commerce could strengthen their role in digital payment ecosystems by managing consent and security protocols. Conversely, they must also address concerns related to liability and dispute resolution when AI-driven purchases are contested by customers.
Security, governance, and customer comfort with delegating financial decisions to AI will be critical factors influencing the adoption pace. Analysts suggest that users may readily accept AI recommendations but remain cautious about fully automated transactions involving money. Visa’s model, which preserves issuer approval authority, aims to mitigate these concerns by embedding human control within the process.
Broader Trends in AI Integration
This pilot reflects a wider industry trend where companies are embedding AI deeper into operational workflows that directly impact revenue and customer interactions. Beyond traditional roles like customer service and fraud detection, enabling AI to trigger payments represents the next frontier in automation within banking.
DBS’s investment in digital banking infrastructure and AI-driven automation positions it at the forefront of this evolution. The success of such initiatives will depend largely on clear boundaries for AI agent capabilities and customer willingness to entrust financial decisions to software.
Photo credit: Patrick Tomasso
Related Articles
OpenAI’s Greg Brockman Affirms GPT Models Are on the Path to Achieving AGI
Converge Bio Secures $25 Million to Advance AI-Driven Drug Discovery with Support from Top Tech Executives
OpenAI Announces API Retirement of Popular GPT-4o Model by February 2026
MayimFlow Leverages AI to Prevent Costly Data Center Leaks Before They Occur
