AI Agent Completes First Autonomous B2B Transaction Under Visa Protocol

Visa and Lianlian DigiTech have taken a significant step toward autonomous B2B commerce by completing the first live transaction through LoopXPay, an AI agent that handled the entire procurement workflow — from sourcing to payment — within a single, controlled sequence. Announced on July 24, 2026, from Singapore, the milestone demonstrates how agentic AI can streamline buying for small and medium businesses (SMBs) that often lack dedicated procurement teams. In the pilot, LoopXPay identified a product sample need, recommended suppliers, compared options, placed the order, and executed the payment, all while operating under predefined spending limits and approval rules. This closed-loop autonomy, backed by Visa’s newly formalized Trusted Agent Protocol and Agentic Directory, signals the start of a broader transition from human-managed B2B workflows to AI-orchestrated commerce. Lianlian, an AI-native global financial infrastructure provider, designed LoopXPay to function within these trust frameworks, ensuring that every agent-driven action is traceable and bound by corporate governance. By registering with Visa’s Agentic Directory, LoopXPay gains a verified identity that participating businesses can trust, addressing one of the biggest barriers to automated corporate spending: the need for assurance that agents are acting legitimately and within approved parameters. For Visa, the move deepens its push beyond consumer payments into the $140 trillion global B2B payments market. While card-based B2B solutions have grown, true automation has been limited by the complexity of cross-supplier comparison, approvals, and payment reconciliation. An AI agent that integrates Visa’s rails with intelligent decision-making could accelerate the shift from checks and manual invoices to real-time, programmatic settlement. The partnership also underscores how traditional payments networks are evolving into platforms for identity, governance, and value-added services — not just transaction processing. The immediate implications for SMBs are practical: reduced time spent on procurement, fewer errors, and tighter enforcement of budget controls without requiring a dedicated purchasing department. Longer-term, as AI agents become more common, the Agentic Directory could become a critical piece of business infrastructure, akin to a digital corporate registry for autonomous spenders. However, this development raises important questions about liability when an agent makes a poor sourcing decision, the security of the agent’s authentication, and the extent to which financial regulations will need to adapt to non-human economic actors. Visa’s protocol addresses some of these by layering identity verification, transparency, and control — but industry-wide standards are still nascent. The LoopXPay demonstration is a proof of concept that agentic B2B commerce can work within existing financial infrastructure, and it sets the stage for a competitive race among payment providers, fintechs, and enterprise software firms to embed AI agents into corporate spending workflows. We can expect further announcements on scaling the Agentic Directory, additional partnerships with ERP systems, and potentially the inclusion of stablecoins or blockchain-based identity to extend trust globally without relying solely on traditional rails.