Alipay has completed 300 million AI-powered agent payments and is developing new protocols to verify that autonomous AI agents act within user intent, signaling that the future of commerce depends less on faster checkouts and more on trustworthy AI decision-making. The milestone, announced in May 2026, covers 95% of general-purpose agent frameworks and represents a fundamental shift in how payment systems must operate when humans delegate tasks to AI agents rather than completing transactions themselves.

Why Are Payment Systems Becoming the Trust Layer for AI Agents?

For decades, the payments industry focused on one core challenge: helping users complete transactions faster, more securely, and at lower cost. Users moved from desktop checkout pages to mobile QR codes to embedded payment buttons in ride-hailing and food delivery apps. But the AI agent era introduces a completely different problem. Instead of a user manually opening an app, searching for products, and clicking to pay, they might simply tell an AI agent: "Order me an iced Americano nearby with good ratings, under 30 yuan, and delivered within half an hour." The agent then automatically understands preferences, integrates mapping, food delivery, coupons, merchant inventory, and payment interfaces, and completes the entire transaction within authorized limits.

At this point, payment systems must verify far more than account validity or merchant legitimacy. They must answer complex questions about agent behavior: Did the user truly authorize this agent? Does the agent's behavior align with the user's original intent? Are the transaction amount, merchant, category, and time within authorized boundaries? If the agent is compromised by prompt injection attacks, malicious plugins, or fraudulent merchants, can the payment system detect and block it? These questions represent a fundamental expansion of what payment infrastructure must do.

What New Payment Protocols Are Emerging to Support Agent Commerce?

Alipay's response has been comprehensive. The platform now supports AI Pay, AI Receive, Token Pay, and an AI Wallet, building an open technical protocol framework called the Agentic Commerce Trust Protocol (ACT) around these capabilities. The company accelerated product development significantly: in June 2026, Alipay launched "Abao," an AI-powered version of its platform, beginning with invite-only testing. By July 2, Abao completed its next iteration and opened to public testing, introducing 72 intelligent service capabilities across multiple domains.

Alipay is not alone in recognizing this shift. Around the same time, other major payment players announced competing protocols:

• OpenAI and Stripe: Launched Agentic Commerce Protocol (ACP), featuring Instant Checkout within chat interfaces
• Mastercard: Introduced Mastercard Agent Pay for agent-based payment authorization
• Visa: Unveiled Trusted Agent Protocol (TAP) and Intelligent Commerce Connect solutions
• China UnionPay: Released APOP, an Open Protocol Framework for Agent Payments

This convergence suggests that the agentic economy will not make payments disappear; rather, it will elevate the importance of payments. Under the new paradigm of "humans delegating tasks to agents," payment systems will no longer be merely the final step for deducting funds. Instead, they will become the central hub for verifying identity, reconstructing intent, expressing authorization, controlling budgets, allocating responsibility, and building trust.

How Should Organizations Prepare for Agent-Based Commerce?

Organizations preparing for this shift should consider several practical steps to align their payment infrastructure with emerging agent frameworks:

• Audit Authorization Models: Review how your current payment systems handle user consent and authorization. Agent-based commerce requires more granular controls that specify transaction limits, merchant categories, time windows, and spending caps that agents must respect.
• Evaluate Protocol Compatibility: Assess whether your payment infrastructure can integrate with emerging agentic protocols like ACT, ACP, TAP, or APOP. Alipay's support for 95% of general-purpose agent frameworks suggests that protocol compatibility will become a competitive requirement.
• Implement Fraud Detection for Agent Behavior: Traditional fraud detection focuses on account compromise and unauthorized transactions. Agent-based commerce requires detecting when agents are compromised by prompt injection, malicious plugins, or fraudulent merchants, which demands different monitoring approaches.
• Plan for Intent Verification: Build systems that can verify whether an agent's behavior aligns with the user's original intent. This goes beyond transaction verification and requires understanding the context and constraints of the user's original request.

The scale of Alipay's achievement underscores the urgency of this transition. Processing 300 million agent payments while supporting 95% of general-purpose agent frameworks demonstrates that agent-based commerce is not a future possibility but a present reality requiring immediate infrastructure adaptation.

The real significance of Alipay's milestone is not simply that it processed a large number of transactions. Rather, it signals that payment systems are evolving from transaction processors into trust verification engines. As AI agents become the primary interface between users and commerce, the ability to verify agent behavior, reconstruct user intent, and enforce authorization boundaries will determine which payment platforms succeed in the agentic economy.