FREE 3-MINUTE ASSESSMENT

How exposed is your AI agent payment flow?

Answer ten concrete control questions. Get a score, priority gaps and a shareable action plan—without an account.

01

Hard spending capIs the final payable amount checked against a numeric maximum?

02

Currency scopeDoes the mandate explicitly restrict permitted currencies?

03

Merchant bindingMust the final seller match an approved stable merchant ID?

04

Explicit consentCan the system prove current human consent for this purchase?

05

Short expiryDoes delegated purchase authority expire automatically?

06

Intent bindingAre the approved intent and final purchase cryptographically linked?

07

Agent identityIs the acting agent bound to the merchant or credential chain?

08

Durable replay protectionCan the same authority be consumed only once across all regions?

09

Model-independent enforcementAre payment rules enforced outside the language model?

10

Decision evidenceDoes every attempt return stable reasons and an evidence receipt?

What the score measures

Execution-time authority: whether the final amount, merchant, currency, consent, identity and retry still match what a person approved.

What it does not measure

Fraud probability, sanctions exposure, model quality or payment credential security. Those require separate controls.

How to verify the result

Use the live validator to attack a sample mandate, then integrate the same policy decision immediately before payment.