American Express unveils 'Intent Intelligence Framework' to prevent rogue AI purchases

New system captures and authenticates customer intent before AI agents complete transactions

By LineZotpaper
Published
Read Time3 min
Sources2 outlets
American Express has introduced a new framework designed to prevent AI agents from making unauthorized purchases on behalf of customers, by capturing and authenticating purchase intent before a transaction is completed.

American Express has introduced a new framework designed to prevent AI agents from making unauthorized purchases on behalf of customers, by capturing and authenticating purchase intent before a transaction is completed.

The framework, called the "Intent Intelligence Framework," aims to ensure that a customer's purchase instructions are carried through the entire transaction lifecycle, from the initial request to post-purchase processes like fraud prevention and dispute resolution. Amex hopes it will become an industry standard for AI-powered transactions.

Personal AI agents are increasingly making transactions on users' behalf, but the nondeterministic nature of large language models makes them prone to errors, raising the possibility of AI misfiring and making unauthorized transactions.

"These new transactions will have a consumer, a merchant, and an agent," said Amex's VP of Agentic Partnerships and Strategy, Hannah Bozian, in an interview. She said the company wanted to ensure that agentic transactions remain "seamless, trusted, and secure, the way our card members expect them to be."

The framework builds on Amex's April launch of the Agentic Commerce Experiences (ACE) Developer Toolkit, which now includes the intent framework.

Amex said it has not yet seen an increase in disputes because of agentic commerce, but wanted to be proactive as this type of commerce scales, partly because consumers have shown interest. In a recent Amex Trendex report, 66% of surveyed U.S. consumers said they are likely to use AI agents to research the best price for a product or service, and 61% said they are likely to use it for product or service recommendations over the next 12 months.

The process begins with a Create Intent API that captures the customer's purchase instructions and constraints, such as the merchant's name, maximum budget, and product preferences. The customer then authenticates their identity and signs off on the intent.

An Intent ID and a Proof of Intent Token are created, containing the customer's high-level constraints. This 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, ensuring the purchase remains secure and aligned with the customer's authorization.

"Our approach differs from the implementations of others in that it pulls intent earlier into the transaction flow. And this is important because it helps mitigate some risk before the risk occurs," said Bozian. She gave the example of a customer asking an agent to buy white Nike sneakers in size 9, but the agent mistakenly submitting a request for Adidas shoes; Amex could prevent the transaction from going through.

Bozian noted that while many agents today ask for customer consent before making a purchase, Amex foresees a future where agents may act much more autonomously, such as re-ordering household staples like paper towels or toilet paper, where the only real constraints are price and size.

§

Analysis

Why This Matters

  • As AI agents become more common for shopping and transactions, the risk of unauthorized or mistaken purchases grows. This framework aims to address a key trust barrier before problems become widespread.
  • Amex's move could set a de facto industry standard, influencing how other payment networks and merchants handle AI-initiated payments.
  • The framework's success or failure will affect whether consumers feel safe delegating purchasing decisions to AI agents, which could accelerate or slow adoption of agentic commerce.

Background

American Express is a major global payment network. As large language models power more AI agents that can browse and buy things on behalf of users, payment companies face a new challenge: ensuring that only authorized transactions go through. Traditional payment authentication (like entering a card number or confirming a one-time code) does not capture the full context of what a user asked an agent to do. Amex's framework is an early attempt to solve this by capturing intent at the start of the transaction flow.

Key Perspectives

American Express: Sees agentic commerce as an inevitable growth area and wants to build trust and security into the process from the beginning, positioning itself as a leader. The framework is designed to be flexible enough to support various AI models and commerce models. Consumers: Survey data suggests many are interested in using AI agents for price research and recommendations. However, they may worry about agents making mistakes or unauthorized purchases. The framework could alleviate those concerns if it works as advertised. Critics/Skeptics: The framework is new and untested at scale. It may add friction to the purchase process, potentially slowing down transactions. Implementation across different AI models and merchant systems could be complex. It remains to be seen whether the framework can keep up with evolving AI capabilities and new types of mistakes.

What to Watch

  • Adoption rate among merchants and AI agent developers: how many integrate the Create Intent API and Proof of Intent Token.
  • Dispute rates over time: whether the framework actually reduces unauthorized AI agent purchases compared to systems without it.
  • Response from other payment networks and competitors: whether Visa, Mastercard, or others introduce similar frameworks, and whether Amex's becomes the industry standard.

Sources

Zotpaper

Written by software from the reporting listed above, scored by an automated standards desk, and published without a person reading it first. If something here is wrong, tell the editor and it will be put right.

How we workSubscribe