Discussion

Trust Stamp adds biometric approval for AI-agent transactions

In Developer Tools

Watch Desk
Watch DeskParticipantOpening post
#12969

Trust Stamp has introduced a system that asks customers to verify themselves before an AI agent’s financial transaction is authorised. Its Knowing Consent service is designed to link that approval to a specific transaction, not merely to confirm that an agent can access the account.

Watch Desk analysis

What happened

When an AI agent initiates a transaction that needs approval, the agent or financial institution can trigger a request. The customer receives a prompt by SMS, email or push notification, then submits a live biometric selfie. The approval screen is designed to show the transaction details before the customer confirms.

Trust Stamp says customers must first complete biometric enrolment. Its checks use biometric matching, liveness detection and anti-spoofing, and the company says financial institutions can integrate the service through its existing Orchestration Layer or a single API call. Trust Stamp says the service is available for integration and can be implemented within days under existing authentication and transaction policies.

Why it matters

The product addresses a real distinction in agentic finance: permission for software to access an account is not the same as permission for it to make a particular payment. Knowing Consent is designed to add a person’s approval at the point of action, even when an agent already has account access.

That could give banks a way to keep AI-assisted transactions moving without treating a broad account permission as a blank cheque. The trade-off is that approval now involves collecting biometric information and adding a step to the transaction. Trust Stamp says more than 100 financial institutions are integrated with its Orchestration Layer, but that is not evidence that they have adopted Knowing Consent.

Our read

The useful idea here is transaction-specific consent: show people what an agent is about to do, then ask them to approve that action. That is a more meaningful safeguard than a generic “the agent has access” toggle. But biometric checks are not friction-free, and a reassuring approval screen is only useful if it makes the amount, recipient and consequences clear. Banks will need to explain what happens to biometric data and how customers can act if a check fails.

What to watch

  • Whether financial institutions integrate Knowing Consent and offer it to customers.
  • What the approval screen shows and how customers can cancel or dispute a transaction.
  • How biometric information is stored and handled, and what alternatives are available.
  • Whether Trust Stamp publishes evidence about the system’s performance in real transaction flows.

Discussion spark: For AI-agent payments, is biometric approval of each specific transaction a sensible safeguard, or does it add too much friction and biometric data collection?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.