Mastercard and Ant International are advancing agentic commerce in Asia Pacific with two payment milestones demonstrating how artificial intelligence (AI) agents can conduct transactions on behalf of consumers and businesses under verified authorization. The collaboration includes a live consumer purchase through Antom, Ant International’s merchant payment platform, and a completed Singapore pilot of Mastercard Agent Pay for Machines (AP4M). Both initiatives address a central challenge facing AI-driven commerce: ensuring that automated agents have verified identities, clearly defined purchasing permissions and accountable payment controls. The developments signal a shift from experimental AI shopping capabilities toward payment infrastructure designed for real-world agent-initiated transactions.
Verifiable Intent Addresses Consumer Authorization
The consumer demonstration focused on a fundamental issue in AI-powered payments: how merchants can establish that an AI agent is acting according to instructions genuinely approved by the customer.
In the live transaction, an AI agent purchased a cup of coffee using the consumer’s previously captured purchasing intent. Antom supported merchant acceptance, while Mastercard Agent Pay with Verifiable Intent provided a record of what the consumer had authorized.
Mastercard Agent Pay combines network tokenization, authentication and consumer intent capabilities to support agent-initiated transactions.
Its Verifiable Intent technology introduces an additional trust layer designed to make consumer authorization verifiable across participating merchants and payment ecosystem providers.
The framework is open, standards-based and protocol-agnostic. It is designed to work across different agentic commerce protocols rather than requiring every participant to operate within a single AI platform.
Mastercard developed Verifiable Intent in collaboration with Google, aligning the framework with emerging standards such as Google’s Agent Payments Protocol (AP2) and Universal Commerce Protocol (UCP).
This approach could help address concerns around unauthorized purchases, unclear agent permissions and disputes involving AI-initiated transactions.
For merchants, the ability to verify what a customer authorized could become an important requirement as AI assistants take on more purchasing responsibilities.
Agent Pay for Machines Extends Payment Controls to Businesses
The second milestone focuses on business procurement and machine-to-machine payments.
Mastercard’s Agent Pay for Machines (AP4M) is designed to let businesses create credentials for AI agents, establish purchasing permissions and define spending limits before allowing those agents to transact.
Unlike conventional online checkout, where a person typically reviews and approves each purchase, machine-to-machine commerce can involve automated transactions between software systems.
These transactions may include purchasing API access, cloud computing resources, data services and other digitally delivered products.
Mastercard introduced AP4M in June 2026 as infrastructure for machine-driven payments, with industry participants including Ant International, Cloudflare, Adyen, Stripe and Coinbase among its early supporters.
The completed Singapore pilot with Antom tested how these controls work in practice.
An AI agent used AP4M credentials, policy restrictions and payment authorization capabilities to purchase digital services within parameters established by a business.
Each transaction generates an auditable record, while merchants can identify the business represented by the purchasing agent.
This creates a framework for delegating routine purchases without removing organizational oversight.
For finance and procurement teams, the distinction is important. Businesses need the efficiency of automated purchasing but cannot afford to lose control over budgets, approved vendors and transaction accountability.
Antom Connects Agentic Payments With Merchant Acceptance
Ant International’s Antom plays an important role in bringing these capabilities into the merchant ecosystem.
The platform combines payment acceptance and merchant digitalization services, helping businesses process transactions across different markets and payment environments.
Its collaboration with Mastercard extends that role into agentic commerce, where merchants must be prepared to accept purchases initiated by AI systems rather than conventional human-operated checkout sessions.
Chunmao Zhu, VP of Technology and Business at Antom, said the collaboration could help merchants reach customers through AI agents and introduce smaller, flexible services that agents can discover, combine and purchase.
This points toward potential changes in digital commerce models.
For example, AI agents could eventually identify individual digital services, compare their capabilities and purchase specific resources without requiring customers or employees to navigate multiple checkout processes.
Ant International also operates across payments, accounts, foreign exchange, treasury and merchant growth services. Integrating these capabilities with agentic commerce could create additional opportunities for merchants to automate financial and operational workflows.
Asia Pacific Emerges as a Testing Ground for Agentic Commerce
The two milestones highlight an important direction for the regional fintech industry.
Asia Pacific has a diverse digital payments ecosystem, with established mobile payment platforms, international payment networks, digital banks and large technology-led commerce businesses.
For Mastercard, working with Ant International provides an opportunity to test agentic payment infrastructure across both consumer purchases and business transactions.
The Singapore pilot also demonstrates why agentic commerce requires more than AI models capable of completing tasks. Payment networks, merchants and financial institutions need shared mechanisms for identity verification, authorization, transaction monitoring and accountability.
The technologies remain at an early adoption stage. Mastercard’s AP4M offering is still being developed through its pilot and early-access ecosystem rather than being a universally available commercial payment service. Mastercard Global+1
The next challenge will be interoperability: ensuring that different AI agents, payment providers, merchants and financial institutions can recognize and enforce consistent authorization rules.
For the broader fintech market, Mastercard and Ant International’s collaboration signals that trust infrastructure may become as important to agentic commerce as payment processing itself.
Market Landscape
The Mastercard–Ant International collaboration reflects the emergence of agentic commerce as a new segment within digital payments infrastructure.
As AI agents move from providing recommendations to executing transactions, payment providers must address new requirements around identity, authorization, security and accountability.
Four developments are shaping this market:
1. Agentic commerce infrastructure
Payment networks are developing systems that allow AI agents to initiate purchases under predefined permissions. Mastercard Agent Pay is an example of this emerging infrastructure.
2. AI payment authorization
Traditional payment authentication verifies payment credentials and customer identity. Agentic commerce introduces another requirement: establishing whether an AI agent has permission to act on someone’s behalf.
3. Machine-to-machine payments
AP4M targets automated transactions between software systems, including payments for APIs, computing resources and digital services. This could create new commercial opportunities for businesses selling machine-accessible services.
4. Open payment standards and interoperability
Standards-based frameworks such as Verifiable Intent could help different AI platforms, merchants and payment providers establish consistent authorization mechanisms.
| Technology | Primary function |
|---|---|
| Mastercard Agent Pay | Secure AI-initiated consumer payments |
| Verifiable Intent | Verifiable consumer authorization |
| Agent Pay for Machines | Business-controlled machine payments |
| Antom | Merchant payment acceptance and digitalization |
| Network tokenization | Protection of payment credentials |
Competitive implications: Mastercard’s strategy positions payment authorization and trust infrastructure as important components of the emerging agentic commerce ecosystem. The ability to support interoperable AI agents while maintaining transaction controls could become a differentiator for payment networks and fintech infrastructure providers.
Top Insights
- Mastercard and Ant International demonstrated a live AI-initiated coffee purchase using verified consumer instructions and Mastercard Agent Pay.
- Verifiable Intent creates a standards-based trust layer that helps merchants establish what consumers authorized AI agents to purchase.
- The Singapore AP4M pilot demonstrated how businesses can define AI agent credentials, purchasing permissions and transaction spending limits.
- Antom provides merchant acceptance capabilities for consumer and business transactions initiated by AI agents.
- The collaboration highlights the growing importance of identity verification, authorization and accountability in emerging agentic commerce infrastructure.
Get in touch with our fintech experts

