IDEMIA Launches Security for Agentic Payments
IDEMIA introduces payment controls designed to prove when an AI agent is authorised to transact for a consumer.
Agentic commerce creates a new payments problem: proving that software genuinely has permission to spend money on someone's behalf.
What happened
IDEMIA Secure Transactions launched an Agentic Commerce system for payment networks and issuers.
The product combines tokenised payment credentials with FIDO2 authentication, consumer-consent verification and configurable restrictions around merchant, amount, category and time.
The goal is to let an AI agent complete a purchase while giving the payment system evidence that the user authorised the transaction and defined the boundaries within which the agent can act.
Why it matters
Traditional payment security assumes a human is actively completing the transaction.
AI agents break that assumption. A user may give an instruction once, while software searches, selects and buys later.
That creates new questions around delegated authority, revocation, fraud liability and how a merchant knows the software is acting within a customer's instructions.
The bigger picture
Payments infrastructure is beginning to adapt to AI agents as economic actors.
The most important technology may not be the shopping agent itself, but the identity and permission layer underneath it.
If agentic commerce scales, banks, networks and payment processors will need standards for machine-to-machine transactions that preserve the controls built for human commerce.