In trade and commerce, agentic artificial intelligence (AI) is poised for substantial adoption and a broader range of applications across more complex workflows.
However, for this vision to materialize, a robust payment infrastructure is essential, according to a new whitepaper by StraitX, Visa, and the Singapore Fintech Association (SFA). This infrastructure should extend beyond a single payment system, enabling agents to interact with diverse payment methods, including cards, bank transfers, local payment networks, and stablecoin-based settlements.
Released in September 2026, the paper emphasizes the need to upgrade payment infrastructure to support the next evolution of commerce and facilitate advanced AI agents.
According to the industry stakeholders, agentic commerce isn’t just about making AI capable of making payments. It also involves building financial infrastructure that can support an increasingly autonomous and interconnected economy.
The experts highlighted four priorities emerging for the development of such agentic payment infrastructure. First, this infrastructure should enable interoperability, and allow agents to connect to multiple payment and settlement mechanisms rather than designing around a single rail.
Agentic commerce will develop across cards, local payment networks, and bank transfers. Hence, this infrastructure should provide a common infrastructure layer that makes different rails accessible and interoperable, the experts said.
This infrastructure should also support stablecoins. These digital currencies will play a critical role in this emerging infrastructure due to their unique capabilities, such as programmability, rapid settlement capabilities, and suitability for frequent or low-value transactions, they claimed.
Machine-to-machine transactions, where agents are required to pay for data, application programming interfaces (APIs), computing resources, or other digital services as part of ongoing workflows, will particularly benefit from these features.
Stakeholders should also establish clear authority and accountability. This involves developing mechanisms to define agent authorization, the entity on whose behalf it acts, and responsibility determination when agents operate outside these parameters. Furthermore, the autonomous executive must have built-in controls, including access controls, transaction limits, monitoring, and runtime safeguards.
Finally, standards and regulatory frameworks must evolve to accommodate this new payment infrastructure. In particular, existing security, compliance and governance frameworks will need to adapt to account for autonomous financial activity. Greater interoperability in areas such as agent identity, delegated authority and transaction controls will help reduce fragmentation as adoption develops, the experts said.
Agentic commerce in Asia
Although agentic commerce is still a relatively new concept, Asia-Pacific (APAC) is anticipated to emerge as the global leader in this field. According to a 2026 report by Deloitte, the region is set to drive around two-thirds of the world’s new retail sales over the next five years, powered by more than 4.3 billion shoppers, 18 megacities and the world’s fastest-growing middle class.
Currently, only 29% of consumer businesses in APAC report adopting agentic AI. However, this figure is expected to surge to 76% within two years.
Deloitte identifies several agentic AI trends that are already transforming retail in APAC. First, retailers and consumer brands are rapidly embracing AI to support marketing, sales and service across areas like content generation and AI assistants. The technology is now redefining how customer engagement is shaped in real time by drawing on richer data, new interactions and continuous analysis to respond more precisely to individual customer needs.
Secondly, stores in markets such as China and Singapore are rapidly evolving beyond transactional spaces and into environments where intelligence is embedded directly into the shopping experience. Capabilities including electronic shelf-labels, intelligent shelving, digital store operations, and agent-driven workflows are no longer experimental efficiencies, and are instead becoming the foundations of a new physical shopping experience.
Another trend highlighted by Deloitte is the deployment of agentic AI in operations. Leading APAC omnichannel retailers are already utilizing specialized AI agents across functions including forecasting, inventory, pricing, fulfillment and customer services, reflecting how agentic AI is expanding beyond supporting work to actively shaping how work flows across the enterprise.
One relevant example is JD.com, one of China’s largest retailers, which uses AI across its business, including across smart logistics and manufacturing to bring products closer to its customers. The firm has already integrated over 14,000 AI agents, including specialized digital workers such as a Demand Agent, Operations Agent, Customs Agent and Courier Assistant to enhance and automate key steps from manufacturing through delivery. These agents now reportedly handle over 18% of all work tasks across its retail, logistics, healthcare and industrial sourcing functions.
Finally, the fourth and last trend is the advent of shopping agents as prominent manifestation of agentic commerce. In particular, consumers are already using AI to search, compare and evaluate options, and in some cases to complete purchases on their behalf.
As these capabilities mature, Deloitte expects shopping agents to begin moving beyond simple transactions toward ongoing delegation, managing replenishment, subscriptions, budget optimisation and even brand choices over time.
Some industry forecasts anticipate a growing share of activity will be handled by agents in coming years, potentially accounting for up to 25% of global e-commerce sales by 2030.
Featured image: Edited by Fintech News Singapore, based on image by stockboy via Magnific
