Instant Payments Governance: Why the Real Gap Is Architecture, Not Speed

Instant payments do not create a governance problem. They compress the time available to make a decision, which makes an existing architectural problem much easier to see.

For many financial institutions, the hard part is not the speed of FedNow or RTP. It is whether fraud, identity, authority, policy and payment execution can come together quickly enough to influence the transaction before settlement becomes final.

That distinction matters because a payment system can work exactly as designed and still carry a bad instruction. The question is whether the institution can recognize that instruction while there is still time to act.

Why instant payments governance has to reach the payment before settlement

In slower payment environments, some controls could operate after initiation because there was more time before final settlement. Instant payments change the timing. FedNow and RTP settle quickly, and although both have formal processes for requesting and returning funds, a sending institution cannot simply recall a payment after it has settled.

The 2025 C&M Software incident in Brazil illustrates the issue. Attackers used legitimate credentials to gain access through a technology provider connected to the country's payment infrastructure. Police said at least R$800 million was diverted, and BMP said reserve accounts belonging to six financial institutions were accessed. Pix itself was not breached. A trusted connection into the payment infrastructure was exploited.

That is a useful distinction for senior leaders. Security around the rail matters, but so does the institution's ability to evaluate the instruction before it reaches final settlement.

Richer data only helps if it changes the decision

ISO 20022 gives financial institutions richer structured payment information. Agentic systems can add context about the request, the initiator and the conditions around it. Fraud and identity platforms add still more signals.

The problem is not a lack of information. It is fragmentation. When those signals sit in different systems with different owners, an institution can know a great deal about a transaction without having one place where that knowledge becomes a consistent decision.

BIS Project Mandala is a useful public example of the direction of travel. Its compliance-by-design approach moved required checks into pre-validation, before the payment instruction was initiated. The project is cross-border and experimental, but the principle is relevant: controls are most valuable while they can still change what happens next.

The question senior leaders should ask first

Before asking which AI tool to adopt, which rail to add or which fraud platform to replace, I would start with a more basic question: where does your institution decide what is permissible before money moves?

• Who owns the full transaction decision across identity, authority, risk and payment execution?

• Does richer transaction information change the decision, or simply pass through the architecture?

• Can a material risk signal reach the payment before settlement, or mostly improve investigation afterward?

• When a new rail or form of money movement is added, does it extend a common governance architecture or create another silo?

This is why I use the term Real-Time Control Layer™. It describes the architectural responsibility for keeping institutional policy and authority connected to the transaction as the surrounding technology changes.

The goal is not to make every AI agent, fraud tactic or payment type predictable. The goal is to make the institution's own decision standard dependable.

That is the governance challenge underneath instant payments modernization, and it will matter even more as agentic commerce, stablecoins and tokenized deposits become part of the same money-movement environment.

The question worth carrying back to your team is simple: where is that decision being made today?

Want the bigger picture?

This article is part of The Money That Thinks, a five-part series in The Instant Edge looking at what happens when richer payment data, agentic AI, adaptive fraud and instant money movement all collide with financial institution architecture.

In the final issue, Before the Money Moves, the Institution Has to Be Clear, I connect those threads and explore what senior leaders should be thinking about as the environment keeps changing.

→ Read the full article on The Instant Edge