Amex introduces Intent Intelligence Framework to prevent AI agents from making unauthorized purchases
American Express introduced the Intent Intelligence Framework, which captures and authenticates a customer's purchase intent before an AI agent completes a transaction. The framework aims to keep customer instructions intact across the transaction lifecycle and prevent agents from making unauthorized or mistaken purchases. Amex hopes it becomes an industry standard for agentic commerce and is already in partner discussions.
- Amex's Intent Intelligence Framework captures purchase intent via a Create Intent API, then issues an Intent ID and Proof of Intent Token so agents check out with secure credentials instead of card numbers
- Amex VP Hannah Bozian says the framework pulls intent earlier into the transaction flow, letting Amex block mismatches like an agent buying Adidas when a customer ordered white Nike sneakers in size 9
- The framework joins the ACE Developer Toolkit launched in April; Amex reports 66% of surveyed US consumers would use AI agents for price research and 61% for recommendations
- Amex is discussing ACE implementation with payments ecosystem partners and Bozian will detail the framework at TechCrunch Disrupt 2026 on October 13
Entities
Related
Visa, Mastercard, Amex back agent payment authorization standards
As AI agents gain the ability to spend money autonomously, card networks are backing new standards requiring agents to prove authorization and spending limits before transactions. This addresses the failure of traditional bot detection when the bot is the customer.
As AI agents gain the ability to spend money autonomously, card networks are backing new standards requiring agents to prove authorization and spending limits before transactions. This addresses the failure of traditional bot detection when the bot is the customer.
Authoryze: Payment controls for AI agents via MCP
Authoryze is a tool that adds payment controls to AI agents through an MCP server, requiring agents to make purchase requests that can be approved or denied. It solves the problem of giving agents credit cards by providing a safe way to manage purchases, preventing wrong or duplicate buys and fraud. Aimed at developers building AI agents that need to make transactions.
Authoryze is a tool that adds payment controls to AI agents through an MCP server, requiring agents to make purchase requests that can be approved or denied. It solves the problem of giving agents credit cards by providing a safe way to manage purchases, preventing wrong or duplicate buys and fraud. Aimed at developers building AI agents that need to make transactions.
Issuer-Sovereign Agentic Payments
An arxiv paper describes a method for AI-agent card payments in which the issuing bank, rather than an external credential provider or card network, enforces spending rules. The cardholder approves a rule once; the bank's authentication component later generates the card authentication value only when the merchant matches the approved rule. This matters to practitioners building agentic payments because it shifts authorization risk-control to the party carrying the liability.
An arxiv paper describes a method for AI-agent card payments in which the issuing bank, rather than an external credential provider or card network, enforces spending rules. The cardholder approves a rule once; the bank's authentication component later generates the card authentication value only when the merchant matches the approved rule. This matters to practitioners building agentic payments because it shifts authorization risk-control to the party carrying the liability.
Payment infrastructure for AI agents to buy things autonomously
A developer built a payment system that lets AI agents make purchases without human intervention, addressing the problem that merchant checkouts assume human users. The system includes cryptographic agent identity, spending limits, replay protection, settlement, and refunds, and connects to the x402 network on Base Sepolia. It solves the problem of agents being blocked by CAPTCHAs or fraud models while allowing owners to control spending.
A developer built a payment system that lets AI agents make purchases without human intervention, addressing the problem that merchant checkouts assume human users. The system includes cryptographic agent identity, spending limits, replay protection, settlement, and refunds, and connects to the x402 network on Base Sepolia. It solves the problem of agents being blocked by CAPTCHAs or fraud models while allowing owners to control spending.
Paynomos: Policy-based spending controls for AI agents
Paynomos is a tool that gives AI agent owners control over agent spending by setting policies like per-transaction limits, daily caps, and vendor rules. Agents get a payment channel scoped to these rules, and any out-of-policy transaction is flagged for owner approval. It solves the problem of uncontrolled AI agent expenditures.
Paynomos is a tool that gives AI agent owners control over agent spending by setting policies like per-transaction limits, daily caps, and vendor rules. Agents get a payment channel scoped to these rules, and any out-of-policy transaction is flagged for owner approval. It solves the problem of uncontrolled AI agent expenditures.