Visa unveiled its new Visa Agentic Commerce infrastructure (officially called Intelligent Commerce) for Asia Pacific on 12 November. It wasn’t merely a new payment feature — it signalled the emergence of AI Commerce Infrastructure designed to tackle a little-noticed problem: retail sites are being flooded by AI agents, and merchants lack a reliable way to distinguish legitimate shoppers from malicious bots.
With AI-driven traffic to retail sites reported to have surged by 4,700% in a single year (see Adobe Data Insights), Visa’s early‑2026 regional pilots give businesses roughly 14 months to prepare their payment systems, product metadata and customer journeys for a future commerce era in which artificial intelligence handles shopping and transactions on behalf of users.
Why Asia Pacific, why now
Visa’s decision to pilot agentic commerce capabilities in Asia Pacific by early 2026 reflects more than geography — it recognises the region’s leadership in mobile payments adoption and its digitally native consumer behaviour. For many retailers and platforms in APAC, mobile-first payments and rapid innovation mean the market will see machine‑initiated transactions sooner than elsewhere.
Deploying AI commerce infrastructure is a fundamental architectural shift: payment systems and back‑end systems must be designed to handle machine‑speed transactions, cross‑platform handoffs and volumes that outpace human shoppers. That requires changes across product, integration and customer experience design.
“Agentic commerce is transforming the very fabric of online payment transactions, requiring a unified ecosystem to unlock its full potential,” said T.R. Ramachandran, head of products and solutions for Asia Pacific at Visa.
“With Visa Intelligent Commerce and its cornerstone, Trusted Agent Protocol, Visa is connecting consumers, AI agents and merchants through secure, scalable solutions.” In short, the platform aims to let agents act on behalf of customers while preserving merchant visibility and customer authorisation.
According to Adobe Data Insights cited in Visa’s announcement, 85% of consumers who’ve used AI for shopping report improved experiences. Yet that enthusiasm hides a practical problem for retailers: merchants can’t reliably distinguish between legitimate agents acting with consumer consent and sophisticated bots intent on fraud or data scraping. Retailers and companies in APAC should therefore start auditing mobile payment flows and agent‑authentication systems now to be ready for the future of commerce.
The technical architecture behind Agentic Commerce
Visa Intelligent Commerce is built from a set of integrated APIs that cover tokenisation, authentication, payment instructions and transaction signals — effectively creating a new protocol layer for AI Commerce Infrastructure. These components work together so agents can request, authorise and complete transactions while preserving the integrity of payments and merchant systems.
At the centre of the design is the Trusted Agent Protocol. In practice, this uses agent‑specific cryptographic signatures and consent tokens to verify that an AI assistant has genuine commerce intent and valid consumer authorisation. That verification step is intended to address gaps where traditional fraud and security systems — tuned to human behaviour — would otherwise flag legitimate agent activity as suspicious.
Existing fraud detection looks for human anomalies — unusual locations, odd timing or atypical product mixes. By contrast, agent‑driven activity often shows machine‑speed checkouts, simultaneous transactions across multiple merchants and purchasing patterns optimised by algorithms. The Trusted Agent Protocol and transaction signals help distinguish these legitimate agent patterns from malicious automation or data scraping.
Importantly for merchants, the infrastructure is designed to maintain consumer visibility even when an agent intermediates. Practical implementations can surface a hashed customer identifier, consent scope and a merchant‑facing consumer token so businesses retain the customer data needed for marketing, loyalty programmes and personalised service without exposing raw payment credentials.
Visa has emphasised an open, low‑code framework to lower integration barriers. That approach aims to let a wide range of platforms, payment processors and commerce tools interoperate at scale — enabling retailers and product teams to adopt agent‑capable flows without rebuilding core systems from scratch.
The ecosystem emerging around AI payments
Visa’s announced partnerships with Ant International, LG Uplus, Microsoft, Perplexity, Stripe and Tencent illustrate how building AI commerce infrastructure at scale is a collaborative effort. Each partner represents a different node in the emerging ecosystem — from telco and platform reach to search and payment processing — and together they form the integration fabric agents will traverse.
These aren’t conventional payment relationships. Instead, they enable a network where AI agents must authenticate across multiple platforms, access payment credentials securely via tokenised flows, and execute transactions that span services within a single consumer intent. In short: platforms, payment processors and merchant systems will need to interoperate seamlessly to support agent‑driven commerce.
Consider a practical flow: a consumer asks Microsoft’s AI assistant to “plan a weekend in Kuala Lumpur.” The agent could use Perplexity to research options (content and comparison), consult inventory and pricing on merchant platforms, route payment through Stripe and settle over Visa’s network — all while exchanging consent tokens and cryptographic signatures that preserve customer authorisation and merchant visibility.
This orchestration requires infrastructure that supports secure handoffs, trust signals and clear audit trails. It also raises regulatory complexity: different Asia Pacific jurisdictions will approach AI agent authorisation, consumer protection, data residency and cross‑border payments in varied ways. Companies and retailers should map partner roles, confirm data flows and verify compliance requirements as part of any integration plan.
What this means for digital commerce
The shift to AI‑mediated transactions will change fundamental assumptions about online retail. Consumer journeys that used to involve browsing, comparing and clicking “buy” will increasingly happen via conversational instructions to AI assistants, shifting emphasis from attention‑based tactics to algorithmic comparability.
Merchants optimising for human attention spans and click‑through rates must rethink product and marketing strategies: agents evaluate options by score and constraint, not emotion. That means clearer product metadata, consistent inventory and pricing APIs, and content structured for algorithmic comparison to improve conversions when agents make purchase decisions.
Visa’s AI commerce infrastructure also reshapes competitive dynamics. Businesses that integrate early can learn agent‑driven sales flows, preserve customer relationships through consented data sharing, and refine fraud detection for machine‑initiated transactions. Those that delay risk operational gaps when agent adoption reaches a tipping point.
Visa showcased Intelligent Commerce at the Singapore FinTech Festival (12–14 November), providing practical visibility into integration requirements and implementation challenges. Given Visa’s reported 4.8 billion credentials, the platform being piloted in APAC could set patterns for agentic commerce globally — but companies should treat this as strategic preparation rather than inevitable predestination.
Practical checklist for retailers and businesses:
- Audit payments and tokenisation flows to ensure agent compatibility.
- Standardise product metadata and expose inventory/price APIs for agents to compare.
- Review customer data permissions and loyalty integration to preserve trust.
- Test fraud detection with agent‑like traffic patterns and verify signal handling.
- Engage payment providers and platform partners early to plan integrations.
Taking these steps now will help retailers convert early agent‑driven purchases into lasting customer experiences.
The road to 2026
Fourteen months until regional pilots begin may feel like plenty of time, but the technical, operational and strategic work required is substantial. Companies and retailers should prioritise audits of payment infrastructure and integration points to ensure compatibility with agent‑driven flows, and review product and UX design so customer experiences remain coherent when agents act on behalf of users.
Security teams must recalibrate monitoring and fraud detection to separate legitimate agent activity from malicious automation. Key signals to track include agent traffic anomalies, API error and latency spikes, token authentication failures and sudden changes in purchase patterns. Prioritise fixes by impact and effort — for example, token and authentication hardening within 30 days, API and inventory readiness within 90 days, and loyalty/integration testing over 180 days.
Visa’s AI commerce infrastructure is more than a payments product: it establishes a model for future commerce that scales trust between agents, customers and merchants. Treat the APAC pilot as an opportunity to test integrations, validate data flows and refine operational playbooks — the lessons learned now will define how commerce systems operate as agents take on a greater role in purchases and customer relationships.
