American Express introduced the Intent Intelligence Framework this week to prevent unauthorized purchases made by AI agents acting on behalf of customers. The framework aims to authenticate and capture a customer’s purchase intent throughout the transaction lifecycle, ensuring secure and trusted AI-powered transactions, the company said. The announcement follows Amex’s April launch of the Agentic Commerce Experiences developer toolkit, which now incorporates this new framework, according to techcrunch.com.
The framework addresses risks posed by the nondeterministic nature of large language models, which can cause AI agents to misfire and complete unintended transactions. Hannah Bozian, Amex’s VP of Agentic Partnerships and Strategy, told TechCrunch the company designed the framework to maintain seamlessness and security for card members while enabling agentic transactions involving a consumer, merchant, and AI agent. Amex has not yet seen a rise in disputes related to AI-driven commerce but developed the technology proactively.
As AI agents increasingly handle transactions, the Intent Intelligence Framework seeks to set an industry standard for authenticating purchase instructions from initiation through completion. This move positions Amex among early adopters addressing AI risks in financial services. The framework complements the ACE Developer Toolkit, which provides APIs and services to support AI commerce, reflecting growing industry focus on integrating AI responsibly in payments and retail sectors.
Amex’s Intent Intelligence Framework is now part of its ACE Developer Toolkit, available to developers building AI-powered commerce solutions. The company’s proactive approach aims to safeguard card members and merchants as AI agents become more prevalent in transaction processing, according to techcrunch.com.