Agentic commerce’s slow burn for stripe, visa, mastercard

Stripe is spearheading the development of programmable, AI-driven payment infrastructure tailored for autonomous AI agents, exemplified by innovations such as the Agent-to-Commerce Protocol and a shared payment token system embedding spend limits and seller-specific controls. These technologies enable safe, seamless, and deterministic transactions in the emerging robot-driven economy, with Stripe’s Machine Payments Protocol allowing AI agents to autonomously manage real budgets and execute payments without traditional web interfaces. The company’s integration of stablecoins via the Tempo blockchain and partnerships with Meta, Google, and Microsoft further underscore its commitment to building a robust, scalable ecosystem for agentic commerce, transforming money into programmable data that moves at software speed.

Legacy payment giants Visa, Mastercard, and American Express are actively adapting their extensive digital credential ecosystems and trust frameworks to the agentic commerce paradigm. Visa leverages its vast network of five billion credentials and 200 million merchants to provide broad-based acceptance and fraud protection, while Mastercard’s Agent Pay and Agent Pay for Machines platforms establish critical liability, authorization, and dispute resolution mechanisms to ensure secure AI-driven transactions. American Express enhances security through decades of AI-powered fraud detection within its closed-loop network, collectively demonstrating that established players remain vital in underpinning trust and governance in this rapidly evolving $1 trillion market.

Addressing the fragmentation and scalability challenges inherent in agentic commerce, Adyen has launched Adyen Agentic, a universal translator middleware that enables merchants to integrate once and transact seamlessly across multiple conversational AI platforms. This open abstraction layer tackles critical bottlenecks such as protocol fragmentation, legacy data structures, and the trust paradox by translating transaction parameters between legacy systems and AI-driven platforms without requiring costly system overhauls. By updating developer documentation to a native agent-ready markdown format using the llms.txt standard, Adyen significantly reduces integration friction, positioning itself as a key enabler of interoperable, scalable agentic commerce infrastructure.

Despite massive investments and high-profile launches by industry giants such as Visa, Stripe, Amazon, and OpenAI, consumer adoption of agentic commerce remains slow and cautious in 2026. While approximately 30% of consumers use AI tools primarily for product discovery and comparison, full autonomous purchasing is still rare, with adoption hovering at early stages of agentic autonomy as Stripe notes we are "hovering on the edges" of levels 1 and 2 rather than the fully autonomous levels 4 and 5 originally envisioned. This gradual uptake reflects persistent behavioral barriers and trust issues, as many consumers remain uncomfortable delegating complete shopping decisions to AI agents.

Merchant hesitancy to relinquish control over the checkout experience and the complexity of integrating competing protocols—such as Google’s Universal Commerce Protocol alongside alternatives like A2A and AP2—further slow the market's momentum. This reluctance stems from concerns over ceding control and the fragmented nature of the emerging AI-driven payment ecosystem, which complicates seamless adoption and contributes to consumer uncertainty. As a result, agentic commerce is currently more of an augmentation to existing shopping experiences rather than a full disruption, with merchants and consumers alike proceeding cautiously.

The slow and measured consumer uptake echoes patterns seen in previous technology cycles like IoT and NFTs, where hype outpaced practical adoption. The broad and sometimes loose application of the term 'agentic commerce' by various companies has diluted its meaning, making it harder for consumers to understand and trust these AI-driven solutions. Consequently, despite the $1 trillion retail and tokenization market race fueled by bold bets from Visa, Stripe, and Amazon, fundamental behavioral and technological challenges continue to reshape the competitive landscape, underscoring that widespread agentic commerce adoption remains an evolving journey rather than an imminent reality.

Despite AI's growing role in product discovery, a significant trust gap remains as consumers hesitate to allow AI agents autonomous spending authority, underscoring critical challenges in identity verification and authorization models. Visa highlights this divide, noting consumers use AI for research but rarely permit AI to execute purchases, while experts like Tim Joslyn emphasize the necessity of recognizing agent payments distinctly and embedding robust trust frameworks to securely authorize these transactions.

Legacy payment networks such as Visa, Mastercard, and American Express are leveraging their established trust infrastructures—comprising clear liability rules, chargeback protections, and decades of AI-driven fraud detection—to address governance challenges in agentic commerce. Mastercard’s 'Agent Pay' initiative exemplifies this approach by authenticating agents, managing liability, and providing dispute resolution pathways, thereby ensuring that autonomous agents are legitimate actors rather than rogue entities, a complexity highlighted by CEO Michael Miebach.

The rise of agentic commerce demands a fundamental rethinking of authorization and governance, shifting from momentary human approvals to delegated, scoped, and time-bounded permissions for software agents operating over fragmented, asynchronous networks. According to industry analyses and the CSA, this evolution requires new trust anchors—such as Google’s cryptographically signed 'Mandates'—and comprehensive identity frameworks that verify legitimacy, enable behavioral monitoring, and maintain auditable logs to balance innovation with security and compliance at scale.

Security risks intensify as AI agents require broad access to sensitive data—including account balances, invoices, and merchant credentials—raising concerns over data leakage, excessive permissions, and unauthorized access. To mitigate these threats, financial institutions must implement sophisticated security architectures incorporating tool authorization, inter-agent communication controls, human oversight, and automated credential rotation, thereby safeguarding agentic payments within a rapidly evolving autonomous economy that challenges traditional governance and trust paradigms.