Personal AI agents are increasingly transacting on users' behalf, but because large language models are nondeterministic, these agents can sometimes act outside their intended instructions — potentially initiating unauthorized purchases. American Express is launching a framework intended to capture and authenticate a customer's purchase intent before an AI agent completes a transaction.
The company says the goal is to ensure the customer's instructions persist throughout the transaction lifecycle — from the initial purchase through subsequent steps. Branded as the "Intent Intelligence Framework," American Express hopes it will serve as an industry standard for AI‑powered transactions. Hannah Bozian, American Express's VP of Agentic Partnerships and Strategy, described the approach in an interview with TechCrunch, saying the firm wants agentic transactions to remain "seamless, trusted, and secure, the way our card members expect them to be."
Integration with ACE developer tools
The framework follows American Express's April rollout of Agentic Commerce Experiences (ACE), a developer toolkit with services for agentic commerce. The Intent Intelligence Framework is now part of ACE; Amex has said it will continue to develop ACE as the agentic commerce ecosystem matures.
How the framework works
The system starts with a Create Intent API that records a customer's purchase instructions and any constraints included in the request — for example, the merchant name, maximum budget, or product preferences. The method of collecting intent may vary with the AI model used, but the framework is intended to be flexible enough to support different agentic commerce models.
After the intent is captured, the customer authenticates their identity and signs off on the recorded intent. The system then issues an Intent ID — a unique identifier for the intent — and a Proof of Intent Token containing the customer's high‑level constraints.
That token is used to obtain a secure digital payment credential that the agent uses at checkout instead of the customer's actual credit card number. According to Amex, this keeps purchases secure while ensuring they align with what the customer authorized the agent to do.
Risk mitigation, disputes, and autonomous reorders
Amex says the Intent ID and token can be used at various stages of the transaction process to support fraud prevention and dispute resolution. Bozian emphasized that Amex's approach differs from others by bringing intent into the transaction flow earlier, which can reduce risk before it materializes.
She offered the example of a customer asking an agent to buy white Nike sneakers in size 9, while the agent mistakenly submits an order for Adidas shoes. In that scenario, Amex's system could prevent the transaction because it wouldn't match the approved intent.
Bozian also noted that while many agents today request customer consent before purchases, agents may become considerably more autonomous in the future. Consumers might permit agents to autonomously reorder household staples (such as paper towels or toilet paper) where the main constraints are price and size rather than merchant or brand. In those cases, having a technological check helps ensure transactions remain aligned with customer intent.
Maintaining a log of which agent authorized what and when could also assist in investigating agent‑related fraud.
Market response and next steps
Amex said it has not yet seen an increase in disputes tied to agentic commerce, but is building the technology proactively as the practice scales. In an Amex Trendex report cited by the company, 66% of surveyed U.S. consumers said they were likely to use AI agents to research the best price for a product or service, and 61% said they were likely to use an AI agent for product or service recommendations over the next 12 months.
American Express is engaged in conversations with partners across the payments ecosystem to implement ACE, including agentic tokens that contain Amex payment credentials and other parts of the developer kit. The company plans to continue active development of ACE and the Intent Intelligence Framework as the agentic commerce space evolves.
Upcoming presentation
Hannah Bozian will discuss the Intent Intelligence Framework and its implications for agentic commerce on October 13 at TechCrunch Disrupt 2026's Industry Stage.
