American Express has published a comprehensive new framework designed to help card issuers verify purchases made by artificial intelligence-powered shopping agents, marking a significant step forward in the rapidly evolving landscape of automated e-commerce.

The payments titan released its Intent Intelligence framework, outlining a standardized method to record a shopper’s precise purchase instructions before automated AI agents execute transactions on their behalf. As autonomous shopping tools gain traction across the digital marketplace, the system aims to provide a reliable audit trail that card issuers can leverage to detect fraud, approve complex payments, and seamlessly resolve disputes when transactions deviate from user expectations. For retailers, the framework offers a vital mechanism to determine whether ordering errors originated from an AI agent’s misinterpretation rather than merchant error.

While American Express confirmed that Intent Intelligence does not yet feature live integrations, the publication of the framework is intended to encourage other payment networks and technology providers to adopt a collaborative approach rather than introduce immediate proprietary capabilities.

“Providing authorized purchase intent before an agent transacts can give issuers greater confidence that an agent is acting as a customer intended, while also helping issuers proactively manage risk and reduce the chance of costly disputes across the payments ecosystem,” said Luke Gebb, head of global innovation at American Express, in the company’s official announcement.

The rollout arrives as financial institutions and fintech developers race to build robust infrastructure for agentic commerce. In this emerging paradigm, AI agents transcend simple search functions, actively browsing catalogs, comparing prices, evaluating product specifications, and completing financial transactions independently.

Despite the technological enthusiasm, consumers remain apprehensive about handing over financial control to autonomous systems. According to a consumer survey of 1,000 U.S. adults released in August by payments company NMI, 58% of respondents expressed deep concern that an AI shopping agent might purchase the wrong item, size, color, or brand. Furthermore, 60% worried that an autonomous agent could easily exceed their intended budget.

Although American Express initially introduced Intent Intelligence earlier in the year, the latest public release makes its technical specifications widely accessible for industry-wide adoption. The company emphasized that its primary objective is to help advance a common industry standard for intent in agentic commerce.

What the New American Express Framework Means for Retailers

For online merchants and traditional retailers alike, AI purchasing mistakes introduce an entirely unprecedented operational challenge. Historically, a merchant fulfilling an order exactly as transmitted could rely on standard processing rules. In the age of autonomous shopping, however, a merchant might process and fulfill an order flawlessly, only to face a contentious chargeback or dispute because the AI agent purchased an item the human shopper never actually authorized.

This exact vulnerability underscores the value of maintaining an immutable record of the shopper’s original instructions, commonly referred to as purchase intent. By documenting clear parameters approved directly by the customer, the framework establishes objective evidence regarding whether an autonomous agent faithfully executed its assigned task or veered off course.

Intent Intelligence serves as a foundational component within the broader Agentic Commerce Experiences developer kit introduced by American Express. The company highlights that its framework diverges from existing implementations by transmitting a customer’s authorized purchase intent to the card issuer significantly earlier in the transaction lifecycle—specifically, before the agent initiates the actual execution of the purchase.

To accommodate nuanced consumer behavior, Intent Intelligence supports an array of advanced purchase controls. Shoppers retain the ability to establish strict maximum spending limits, expiration dates, and preferred merchant lists. Furthermore, consumers can program temporal conditions, such as instructing an agent to complete a purchase only when a heavily discounted item returns to stock or when a specific shopping holiday begins.

The technical architecture also accounts for multifaceted and conditional purchasing scenarios. For instance, a consumer might authorize an AI agent to purchase travel insurance strictly conditional upon the successful booking of the corresponding airline tickets. The system is equally equipped to manage fragmented transactions across multiple distinct platforms, such as coordinating the purchase of footwear from one online retailer and matching socks from an entirely separate merchant.

How American Express Intent Intelligence Works

Designed to operate seamlessly across diverse agentic commerce models, Intent Intelligence functions independently of whichever entity initially captures the shopper’s instructions. Depending on the architectural design of a given shopping agent, instructions can be transmitted straight to a card issuer, routed through a digital wallet provider, or managed via alternative technical intermediaries.

By standardizing this data flow, the framework ensures that payment verifications remain consistent regardless of the underlying software interface utilized by the consumer.

Why AI Shopping Errors Create New Payment Risks

Beyond empowering card issuers with verification tools, the framework aligns closely with broader strategies developed by American Express to mitigate consumer and merchant liabilities stemming from automated errors.

When American Express initially debuted its developer kit, it outlined plans for a specialized coverage tier termed Agent Purchase Protection. This anticipated protection is designed to cover eligible transactions executed by registered AI agents that successfully utilize the developer kit and authenticate purchase instructions through the framework. For merchants, the ultimate promise of this coverage is a meaningful reduction in costly chargebacks and administrative disputes.

Nevertheless, American Express confirmed that these protections are not yet active, as live integrations utilizing Intent Intelligence are still in development. Until these systems go live, cardholders maintain standard protections and benefits associated with their existing American Express accounts.

This transitional phase introduces complex questions regarding accountability when automated shopping goes awry. In NMI’s consumer survey, 34% of respondents stated they would hold retailers directly responsible if an AI tool executed an unauthorized purchase prematurely. Another 24% indicated they would blame their payment provider, while a substantial 62% placed primary responsibility on the AI technology provider itself.

Where American Express Fits in Agentic Commerce

American Express is far from alone in constructing secure bridges for AI-driven commerce. Major financial institutions and technology giants have rapidly deployed proprietary solutions to capture the future of automated transactions.

In the spring of the previous year, Visa and Mastercard launched dedicated initiatives enabling developers to securely link conversational AI agents directly into their respective global payment networks. Visa’s Intelligent Commerce program provides developers with specialized application programming interfaces to embed payment functions into AI tools, supported by major technology partners including OpenAI, Microsoft, Anthropic, Stripe, and Samsung. Concurrently, Mastercard introduced Agent Pay, an infrastructure designed to facilitate autonomous consumer and business-to-business transactions through conversational platforms.

Simultaneously, leading technology enterprises have advanced independent open standards. OpenAI and Stripe co-developed the Agentic Commerce Protocol to facilitate secure merchant connections for AI checkouts. Google countered with its Agent Payments Protocol, which utilizes cryptographically signed records to capture user consent, drawing early adoption from platforms such as Salesforce, Shopify, Cloudflare, and Etsy.

Google subsequently introduced its Universal Commerce Protocol in collaboration with Shopify, establishing an open standard that major retailers like Newegg, Wayfair, Walmart, and Macy’s have integrated to allow direct checkouts within conversational AI environments like Gemini. As these disparate protocols and frameworks mature, the digital payments ecosystem continues to move toward a unified standard for autonomous transactions.