Mandate-checked initiation
Each agent payment is checked against its mandate and limits before it is sent to the bank.
Payments
Payment service providers
Payment initiation was built for a customer at a screen, not for an agent acting later.
Payment initiation and account information services assume a customer who authenticates at the moment of the payment. Agents act on their behalf and often when the customer is not present, which raises questions about authentication, exemptions and liability under PSD2, the UK Payment Services Regulations 2017 and their equivalents. AgenticOpenFinance™ lets a provider check the agent, the mandate and the policy, and bring the customer in only when the rules require it.
What agents do
What you build on AgenticOpenFinance™
Decoupled approval
Where authentication is needed, the customer approves on their own device and the approval is bound to the amount and payee.
Rail routing
The provider routes each payment to the rail the mandate allows, with a fallback and timestamps for each step.
Idempotent writes
Retries by agents never create duplicate payments; a replayed key returns the original result.
Evidence for liability
Each payment carries the evidence a provider needs when a customer later claims it was unauthorised.
Value chain
The actors in this chain
AgenticOpenFinance™ connects them with one identity check, one mandate model, one policy and one evidence trail.
CustomerAgent operatorPayment initiation providerAccount servicing bankPayeeSupervisor
Typical flow
A governed agent flow, step by step
- Agent requests payment
- Signature and mandate check
- Policy verdict
- Customer approval if required
- Initiation at the bank
- Status and settlement
- Evidence sealed
Live demo
See it in the terminal
The treasury sweep agent asks to move £48,000; above the £20,000 threshold the finance lead decides (demo). Demo data
DemoAll companies, people, agents and amounts are fictional. Protocol names refer to public specifications; naming one is not a partnership or an endorsement.