Priya Jaiswal is a recognized authority in the global financial landscape, bringing years of expertise in banking, business, and international market analysis to the table. As Visa launches its “Agentic Ready” program, Jaiswal provides a seasoned perspective on the transition from manual transactions to a world where AI agents negotiate and shop on our behalf. In this discussion, we explore the operational shifts, regional regulatory landscapes, and technical frameworks like tokenization that are paving the way for autonomous commerce.
How does the shift toward AI agents autonomously discovering and completing purchases change the traditional consumer journey, and what specific technical hurdles must issuing partners overcome to ensure these transactions remain secure and scalable?
The shift to agentic commerce represents a fundamental reimagining of the shopping experience where the human moves from an active searcher to a high-level supervisor. Instead of a person manually scrolling through marketplaces, these AI agents are now designed to autonomously discover, negotiate, and complete purchases, which creates a massive technical demand for issuing partners. To ensure these transactions remain secure, partners like Commerzbank and Nexi Group must validate new pathways for agent-initiated transactions within controlled production environments. It is a high-stakes balancing act where the infrastructure must handle machine-speed decisions while maintaining the strict protections and network-level trust that consumers expect from their financial institutions.
Major institutions like Revolut, HSBC UK, and Barclays are currently testing agent-initiated payments in controlled environments. What criteria define a successful production test for these partners, and how do they balance autonomous machine efficiency with the need for user-controlled oversight?
For institutions like Revolut, HSBC UK, and Barclays, a successful production test is defined by the seamless execution of a transaction without compromising the user’s sense of security. These partners are working in close collaboration with merchants to see if agentic platforms can manage the full lifecycle of a payment while maintaining the “trust and control” that underpins the network. Balancing efficiency with oversight requires a structured pathway where the issuer provides a transparent safety net that monitors AI behavior in real-time. The goal is to reach a point where the speed of the machine does not outpace the regulatory and emotional requirements of the human account holder.
Agentic commerce relies heavily on tokenization, identity verification, and specialized APIs to function within existing networks. How do these underlying technologies mitigate fraud risks during autonomous transactions, and what specific steps are required to integrate these protocols into legacy banking systems?
At the heart of this transition is the framework first developed under the Visa Intelligent Commerce initiative, which launched in April 2025 to provide the necessary architecture for machine-led payments. We are talking about the deep integration of tokenization, identity verification, and specialized network-level APIs that act as a digital credential for the AI agent. These technologies mitigate fraud by ensuring that every “handshake” between the agent and the merchant is authenticated through multiple layers of risk controls before any funds move. For legacy banking systems, this integration requires a sophisticated upgrade to existing payment flow controls, allowing these autonomous flows to sync perfectly with traditional back-end infrastructure.
While initial phases are targeting Europe and the UK, earlier testing occurred in regions like the UAE and Singapore. What regional regulatory differences have surfaced during these cross-border trials, and how can a unified global standard for AI-led payments be established despite varying local compliance requirements?
Moving from the initial phases in the UK and Europe into regions like Singapore and the UAE has revealed that while the technology is universal, the regulatory appetite for autonomous finance varies by territory. Earlier trials with Aldar in the UAE and DBS Bank in Singapore, along with Santander’s recent tests in Latin America, showed that local compliance frameworks often dictate the specific level of autonomy an AI agent can possess. Establishing a unified global standard requires building on infrastructure people already trust while harmonizing these regional nuances into a single, scalable protocol. It is about creating a common language for AI agents so that a purchase initiated in London feels as secure and compliant as one completed in Singapore.
What is your forecast for agentic commerce?
I believe we are entering an era where “buying” becomes a background process handled by specialized digital assistants rather than a manual, time-consuming chore. Within the next few years, the infrastructure being refined today by the Agentic Ready program will likely become the standard for retail banking across the globe. We will see a shift where consumers focus entirely on their preferences and lifestyle values, while AI agents handle the cold, hard numbers of negotiation and settlement. It is a future where commerce is always “on,” yet inherently more secure because the human element of error is replaced by high-fidelity, tokenized automation.
