Card networks, processors, and banks building standards, infrastructure, and trust layers for agent-initiated payments

Payments Networks and Agentic Standards

The landscape of agent-initiated payments continues to evolve rapidly as card networks, processors, banks, and fintech innovators deepen their collaboration to build the necessary standards, infrastructure, and trust frameworks. These efforts aim to enable autonomous AI agents to execute payments securely, compliantly, and at scale—transforming how commerce is conducted across industries and geographies.

Expanding Pilots and Industry Standards Pave the Way for Agentic Payments

Over the past year, significant progress has been made in piloting agent-initiated payments and defining open standards that facilitate trustworthy, programmable AI-driven commerce:

Mastercard’s Agent Pay Framework remains a cornerstone, having enabled the first fully regulated, end-to-end autonomous payment with Santander in Europe. This pilot demonstrated critical capabilities such as agent identity authentication, consumer authorization confirmation, and transaction integrity validation within a strict regulatory environment.
Visa’s Intelligent Commerce initiative, now collaborating with DBS Bank in Malaysia, is pushing boundaries around cultural and regulatory nuance, including Shariah-compliant agent-initiated payment flows. This broadens the applicability of agentic commerce across diverse markets.
Nexi Group has advanced the Model Context Protocol (MCP), which standardizes AI agent interactions with payment infrastructure, enabling more complex and regulated workflows, especially in B2B and heavily regulated sectors.
Stripe’s innovation around Shared Payment Tokens (SPTs) has empowered AI agents to autonomously negotiate payment terms, including flexible financing solutions such as Buy Now, Pay Later (BNPL). Their exploration of blockchain-based settlements via the Tempo blockchain reflects a forward-looking approach to handling programmable, agentic payment settlements in real time.
Spreedly’s introduction of a “Live Channel” specifically tailored for agentic commerce payments offers dynamic escrow and BNPL orchestration, simplifying merchant operations and reducing costs associated with agent-initiated transactions.
Collaboration between Visa, Mastercard, Stripe, and Google continues to harmonize interoperable standards across authorization, settlement, and fraud prevention. This joint effort is critical to creating a unified ecosystem that supports cross-platform agentic payments with consistent security and compliance.

Trust, Verification, and Settlement: The Backbone of Secure Agentic Commerce

Building trust in autonomous payment execution requires robust layers of verification, provenance, and fraud prevention:

Mastercard’s open standard for AI agent transaction verification is a pioneering development addressing three central questions for every agent-initiated payment:
1. Did the consumer explicitly authorize the purchase?
2. Is the AI agent’s identity verified and trustworthy?
3. Has the transaction remained intact and free from tampering?
This framework aims to close liability gaps and mitigate fraud vectors unique to autonomous agents.
The Center for AI Standards and Innovation (CAISI) at NIST is advancing cryptographic identity verification, behavioral authentication, and provenance standards tailored for AI commerce. Their work ensures alignment with GDPR, PSD2, PCI-DSS, and emerging AI-specific regulatory frameworks.
Real-time tamper-proof logging, manifest validation, and automated metadata provenance tools have become essential. For instance, Nextwaves’ UCP Checker Tool enables merchants to detect and remediate manifest errors that could undermine AI agent trust and degrade customer experience.
Experts emphasize the need for adaptive, AI-driven fraud prevention that models autonomous agent behaviors and detects novel attack patterns. Donald Kossmann, CTO of Chargebacks911, highlights:

“Autonomous AI transactions present novel behavioral patterns and attack vectors that traditional fraud systems are ill-equipped to detect.”
Traditional settlement rails are being augmented by blockchain innovations. Stripe’s Tempo blockchain experiments aim to offer transparent, real-time global settlement tailored specifically for programmable agentic payments, potentially reducing friction and settlement times.

Real-World Pilots Validate Compliance and Cultural Adaptation

Concrete pilots continue to validate the practical viability of agent-initiated payments across various contexts:

The Santander–Mastercard Autonomous Payment Pilot in Europe set an industry precedent as the first fully regulated, agentic payment completed end-to-end.
DBS Bank and Visa’s Intelligent Commerce trial in Malaysia demonstrated cultural and regulatory adaptability by incorporating Shariah-compliant payment flows.
Australia’s Mastercard Agent Pay launch with Commonwealth Bank of Australia (CBA) reflects growing international momentum and maturity of agentic payment frameworks.

These pilots serve as critical testbeds to refine standards, workflows, and compliance mechanisms, while fostering trust among merchants and consumers.

Platform Dynamics and Emerging Market Realities

Despite technical and regulatory advances, the deployment of agentic commerce faces evolving platform dynamics that impact its pace and scope:

OpenAI’s recent decision to quietly pull back on enabling users to run e-commerce transactions directly within ChatGPT highlights the cautious approach platforms are taking. Concerns around regulatory compliance, fraud risks, and consumer protection have led to a more measured rollout of direct agentic commerce capabilities.
This shift underscores the importance of interoperable standards and robust trust frameworks to enable safe, scalable agent-initiated payments across diverse platforms and ecosystems.

What Stakeholders Must Do Next

To harness the efficiencies and innovation of agentic commerce, stakeholders across the payments ecosystem must:

Engage proactively with emerging standards such as Mastercard’s AI agent transaction verification, Nexi’s MCP, and Stripe’s Shared Payment Tokens.
Participate in pilot programs and sandbox environments to test and refine agentic payment workflows, ensuring compliance and operational readiness.
Integrate advanced verification, provenance, and adaptive fraud detection tools to build consumer and merchant confidence.
Collaborate across industry consortia and regulatory bodies to harmonize frameworks that support the unique demands of autonomous AI payments.

Conclusion: Towards an Autonomous, Trusted Payments Future

The convergence of card networks, processors, banks, and fintech innovators is steadily transforming agent-initiated payments from experimental pilots into operational realities. The combination of open standards, trust layers, and cutting-edge settlement infrastructure is enabling AI agents to autonomously and securely execute payments with confidence across diverse markets.

While emerging platform caution—exemplified by OpenAI’s pullback—signals the complexity ahead, the industry’s collaborative momentum toward interoperable, secure, and compliant agentic commerce remains strong. As Mastercard, Santander, DBS, Nexi, Stripe, and others continue to push boundaries, the future of commerce is increasingly autonomous, trusted, and agent-executed—ushering in a new paradigm for payments at scale.

Sources (18)