American Express has introduced a new technological framework designed to help card issuers verify purchases made by artificial intelligence-powered shopping agents, marking a significant step forward in the evolving landscape of automated retail transactions.

The payments giant released its Intent Intelligence framework, outlining a standardized method for recording a shopper’s specific purchase instructions before AI systems execute transactions on their behalf. As autonomous shopping tools gain traction across the digital marketplace, financial institutions and merchants face mounting challenges regarding transaction verification, liability, and consumer trust.

Under the newly published framework, card issuers can utilize documented purchase instructions to detect potential fraud, streamline payment approvals, and resolve transaction disputes when outcomes deviate from expectations. For retailers, the system offers a reliable mechanism to determine whether an ordering error originated from an automated agent misunderstanding a prompt or from a fulfillment mistake on the merchant’s end.

While American Express confirmed that the Intent Intelligence system does not yet feature live integrations, the publication of the framework is intended to encourage broader adoption across the payment and technology sectors. Rather than rolling out an immediate commercial capability, the company aims to establish a collaborative industry approach for managing consumer intent in automated commerce.

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, according to Luke Gebb, head of global innovation at American Express.

The strategic move arrives as major financial institutions and technology companies race to develop infrastructure for agentic commerce. This emerging retail paradigm allows autonomous AI agents to search for products, evaluate pricing, negotiate terms, and complete checkouts independently on behalf of consumers.

Despite the technological momentum, mainstream consumers remain notably cautious about surrendering purchasing control to autonomous software. Industry research highlights pervasive consumer anxieties surrounding automated shopping tools. In a comprehensive survey of 1,000 U.S. adults released in August by payments company NMI, 58% of respondents expressed worry that an AI shopping agent might purchase the incorrect item, size, color, or brand. Furthermore, 60% of surveyed consumers voiced concern that an autonomous agent could easily overspend or exceed their designated financial limits.

Although American Express initially introduced Intent Intelligence as part of a broader initiative in April, the latest release makes the technical architecture publicly available for other industry players to build upon. The overarching objective is to help foster a common, interoperable industry approach for managing purchase intent as agentic commerce expands.

What the new American Express framework means for retailers

For merchants and digital retailers, the rise of AI-driven purchasing errors presents an unprecedented operational challenge. A merchant might fulfill an incoming order precisely as transmitted by an automated system, only to face a subsequent chargeback or dispute because the AI agent purchased an item the consumer never actually requested or authorized.

This is precisely where a formalized, verifiable record of the shopper’s original instructions—referred to as purchase intent—becomes vital. By documenting parameters approved directly by the customer, the framework provides clear evidentiary data regarding whether an autonomous agent strictly followed user commands.

Intent Intelligence is one of several specialized services originally introduced by American Express in April through its Agentic Commerce Experiences (ACE) developer kit. Company representatives noted that the framework differentiates itself from existing market implementations by transmitting a customer’s authorized purchase intent to the card issuer much earlier in the transaction lifecycle, well before an autonomous agent executes the final payment.

The framework supports a wide variety of granular purchase controls. Shoppers can establish strict maximum spending limits, set strict expiration dates for purchasing windows, and designate preferred merchants. Additionally, consumers can specify precise temporal conditions for when an agent should execute a transaction, such as waiting until a sold-out item returns to inventory or automatically capitalizing on Cyber Monday promotional pricing.

The system is also engineered to accommodate multi-step, complex purchasing scenarios. For instance, a consumer could authorize an AI agent to purchase travel insurance strictly conditional upon the successful acquisition of accompanying flight tickets. The architecture can similarly manage coordinated, multi-transaction shopping journeys, such as purchasing matching shoes and socks from entirely separate online retailers.

How American Express’s Intent Intelligence framework works

The architecture behind Intent Intelligence is intentionally designed to operate seamlessly across diverse agentic commerce models, regardless of which entity collects the initial instructions from the shopper. Depending on how a specific AI agent is configured to handle financial transactions, the platform can transmit authorization instructions directly to a card issuer, or route them securely through a digital wallet provider or another authorized intermediary.

Beyond helping financial institutions verify transactions, the framework directly supports American Express’s broader strategic vision to protect both consumers and merchants when automated agents make mistakes. When the company initially unveiled the ACE developer kit, it outlined plans for a coverage category known as Agent Purchase Protection.

This specialized protection is intended to apply to eligible transactions involving registered AI agents that utilize the developer kit and successfully transmit authenticated purchase instructions. For merchants, the primary advantage is a substantial reduction in expensive chargebacks and disputed transactions.

However, American Express confirmed that these protections are not yet commercially active, as there are currently no live ACE integrations utilizing Intent Intelligence, which remains a mandatory prerequisite to qualify for the coverage. In the interim, card members retain the standard consumer protections and benefits already associated with their eligible American Express cards.

This evolving layer of technical complexity also influences how consumers attribute responsibility when automated transactions fail. In the NMI survey, 34% of consumers indicated they would hold the retailer directly responsible if an AI tool executed an unwanted purchase prematurely. Another 24% stated they would blame their payment provider, while a commanding 62% reported they would hold the underlying AI technology provider accountable.

Where American Express fits in agentic commerce

American Express is far from alone in developing robust infrastructure to secure and standardize AI-driven shopping. Major payment networks and technology companies have aggressively entered the space to establish dominance over automated transaction rails.

In the spring of 2025, Visa and Mastercard both debuted dedicated offerings allowing developers to integrate AI agents securely into their global payment networks. Visa’s Intelligent Commerce program provides developers with application programming interfaces and development tools to embed payment capabilities directly into conversational and autonomous AI agents. To facilitate a robust global rollout, Visa collaborated closely with prominent technology leaders, including OpenAI, Microsoft, Anthropic, Stripe, and Samsung.

Simultaneously, Mastercard introduced Agent Pay, a technology designed to enable AI agents to complete secure transactions on behalf of users. The platform supports both consumer-facing applications and complex business-to-business environments.

Technology enterprises are also establishing competing open standards and protocols for agentic commerce. OpenAI and Stripe co-developed the Agentic Commerce Protocol, which allows AI agents to interface directly with merchant checkout systems.

Meanwhile, Google introduced the Agent Payments Protocol, which relies on digitally signed records to document consumer authorization, drawing early support from enterprise platforms including Salesforce, Shopify, Cloudflare, and Etsy. Google expanded its footprint with the Universal Commerce Protocol, developed in partnership with Shopify and other industry stakeholders to streamline checkout within conversational AI environments. Major retailers, including Newegg, Wayfair, Walmart, and Macy’s, have since adopted the protocol to enable direct purchasing through AI discovery modes.