Visa completes Greater China’s first AI-powered B2B “agentic” cross-border payment with Lianlian

Visa said it has completed what it describes as Greater China’s first AI powered business-to-business, or B2B, agentic payment. The demonstration paired Visa with Lianlian, according to the report, and focused on moving a cross-border transaction from start to finish using AI to coordinate steps rather than relying on manual orchestration at each stage.

At the center of the test was Visa’s Agentic Directory, a tool Visa uses to help determine where and how payment actions should occur. In this case, the transaction was carried out “end to end” by AI, while still constrained by preset controls set up for the demo, the report says.

The writeup characterizes the result as autonomous within those guardrails, implying the AI system handled the execution flow for the payment journey rather than simply generating instructions for humans to carry out. Visa did not, in the posted report, disclose additional technical specifications such as model architecture, how the preset controls were implemented, or what safeguards were used to limit errors.

The transaction is framed as a step toward broader agentic payment workflows, where software can discover, select, and coordinate services needed for a transaction. For payments networks and intermediaries, the potential value is faster decisioning and more consistent execution across multi-step processes that typically involve routing, messaging, and compliance checks.

Visa and Lianlian did not provide further detail in the published post about the payment corridor (which countries or markets were involved), the settlement method, the transaction size, or whether the demonstration used live production rails or a test environment. The report also did not specify the commercial use case targeted, such as supply-chain finance, collections, payouts, or invoice settlement.

Even so, the description points to an emerging focus in financial technology, where “agentic” systems are used to reduce the need for human intervention in operational workflows. In payments, that can matter because cross-border flows often require coordinated steps that can be difficult to manage consistently at scale, especially as payment types and regulatory requirements vary by corridor.

Sector context matters as well. Visa is one of the payments networks investing in digital infrastructure and, increasingly, in software and services that can help participants move money with fewer operational frictions. Demonstrations that show AI-driven orchestration, if repeated and scaled, can become a pathway for networks to modernize how complex payments are set up and executed.

A key caveat is that this report reads as an announcement of completion rather than a published, independently audited performance analysis. Visa did not disclose metrics such as processing time improvements, failure rates, reconciliation outcomes, or how the system performed under stress conditions in the post described here.

For the next announcement to watch, the market will likely look for follow-on updates from Visa that clarify whether the approach is moving beyond controlled demonstrations, including any published details about corridors, transaction volumes, operational safeguards, and integration timelines with other payments participants in Greater China.