American Express introduces framework to verify AI agent purchases
Tech

American Express introduces framework to verify AI agent purchases

TechNews Editorial
TechNews EditorialOct 13, 2026 · 2 min read
Share

Why it matters

The framework could help Amex check AI agent purchases against customer-approved limits and provide records for fraud prevention and disputes.

The facts

  • American Express introduced a framework to verify customer intent before an AI agent completes a purchase.
  • Customers approve captured instructions, which are linked to a token used to obtain a digital payment credential.
  • Amex says it has not seen more agentic commerce disputes and plans to keep developing its ACE toolkit.

Advertisement

American Express has introduced a framework to capture and authenticate what a customer wants to buy before an AI agent completes a purchase. Called the Intent Intelligence Framework, it is designed to check whether a transaction matches the customer's instructions.

AI agents can make purchases for users, but they can also make mistakes. Amex said it has not seen an increase in disputes tied to agentic commerce. It is developing the framework as this type of shopping grows and hopes it will become an industry standard.

An API records instructions and generates payment tokens

The framework starts with a Create Intent API. It records the customer's instructions and any limits, such as a merchant name, maximum budget or product preferences. How those details are collected may vary by AI model. Amex VP of Agentic Partnerships and Strategy Hannah Bozian said the framework is intended to support different agentic commerce models.

The customer then verifies their identity and approves the captured intent. The system creates an Intent ID and a Proof of Intent Token containing the customer's high-level constraints. The token is used to obtain a secure digital payment credential for checkout, so the agent uses that credential instead of the customer's card number.

Read nextAmazon introduces priming to build hybrid AI models from pretrained Transformers

Amex says the ID and token can also help with fraud prevention and dispute resolution. Bozian gave an example: if a customer asks for white Nike sneakers in size 9 and an agent tries to buy Adidas shoes, Amex could stop the transaction. A record of what a customer authorized, which agent acted and when could also help address fraud involving an agent, she said.

Bozian said agents may eventually make some repeat purchases with less direct customer involvement. For household staples such as paper towels or toilet paper, a customer might set limits on price and size without specifying a brand or merchant. The framework is meant to check those limits during the transaction.

A customer compares a sneaker authorization with a different shoe pictured in a purchase record, pushing the mismatched order aside.
Illustration: AI & Tech News

The tool joins Amex's broader developer toolkit

The framework joins Amex's Agentic Commerce Experiences, or ACE Developer Toolkit, launched in April. Amex said it is discussing ACE with partners across the payments ecosystem. ACE also includes agentic tokens containing Amex payment credentials.

In a recent Amex Trendex report, 66% of surveyed U.S. consumers said they were likely to use AI agents to research the best price for a product or service. Another 61% said they were likely to use one for a recommendation over the next 12 months. Amex said it plans to keep developing ACE as the agentic commerce ecosystem evolves.

Newsletter

Get the best AI & tech news daily

A concise daily digest. Unsubscribe anytime.

We use your email only to send this newsletter.

Keep reading