A payment bot did not go rogue - the architecture did. A timeout produced two valid-looking instructions for the same payout, because a downstream retry never learned the first was already in flight. Mandate-based authorization like AP2 still needs a runtime control plane with real state discipline, not policy in a config file.
An invoice agent hits a transient timeout, retries, and two payments clear to the same vendor. The model calls it success. Ops calls it a week of cleanup. Spend limits and roles are a speed bump. Real delegated authority is a delegation object, policy on every attempt, ledger-bound idempotency, and an instant kill switch.
A wire investigation is the moment confidence breaks - the payment has moved and the clock is running. The Fedwire Funds Service changes planned for November 2026 turn investigation messages into structured workflows and add a Create Payment Return feature. Treat it as a rail feature, not a back-office patch.
The Fed's 2026 pricing notice introduces Payee Name Verification at two cents a transaction, starting on FedACH. It works by comparing a submitted payee name against names previously seen on that account. From an architect's view that is not a yes-or-no check - it is a new control point that reshapes routing, exceptions, and reconciliation.
A payroll rollout taught me the friction is never the integration. It is funding caps, routing permissions, and what can clear directly versus what has to detour through a sponsor bank. That is why the Fed's new payment-account RFI matters - access is becoming a design decision, not just a policy one.
A duplicate API call looks harmless until two payments clear and operations is left explaining the gap. The same error behaves differently on every rail - ACH hides it until batch, RTP and FedNow clear it instantly and final, wire forces a manual recall. On a multi-rail platform, governance is the architecture.