Earlier this month, Sardine and our subsidiary Sonar AI submitted comments to the Office of the Comptroller of the Currency (OCC), the Federal Reserve, and the Federal Deposit Insurance Corporation (FDIC) in response to their Request for Information on Potential Actions to Address Payments Fraud.

Our submission focused on the fact that fraud has evolved into a systemic risk for the U.S. financial system. From check fraud to instant payments scams, fraudsters are exploiting gaps between payment rails, leveraging mule networks, and weaponizing the speed of modern settlement systems.

As regulators gather input on how to strengthen the system, we stressed that traditional tools are no longer enough. Incremental fixes won’t address problems at this scale.

What’s needed is an AI-driven fraud and compliance platform that unifies risk signals, adapts in real time, and empowers institutions of all sizes to act before the money is gone.

The fraud problem has become structural

The innovations that made payments faster and more convenient also outpaced the controls around them. Because risk infrastructure hasn’t evolved at the same speed, it's created the conditions that fraudsters now exploit.

• Faster fraud, slower recovery: Funds availability rules were written for paper checks, when clearing times gave banks a buffer to spot fraud before money went out. Today those same rules govern mobile deposits and real-time withdrawals, which allows criminals to move money before recovery can even start.
• Cross-rail blind spots: Money movement has become multi-rail, yet dispute mechanisms haven’t adapted. An ACH credit can be posted and then instantly withdrawn or wired out. That sequence makes disputes nearly impossible to resolve in time, because the money is gone before the process even begins.
• Fraud & AML disconnect: Institutions run fraud and compliance as two separate functions, yet criminals don’t distinguish between the two. Fraud becomes money laundering the moment stolen funds need to be used or withdrawn. Treating the two separately creates blind spots that allow those flows to continue.
• Fragmented data: Adopting specialized vendors for different parts of the stack has created a fragmentation problem for risk teams. When data is trapped in silos, attribution of mule accounts is delayed and early intervention doesn’t happen.
• AI adoption gap: Fraudsters are already using AI to generate deepfakes, spin up synthetic identities, and run scams at scale. But financial institutions haven’t kept pace. Compliance teams hold back because they worry about how regulators will react, and many risk teams are held back by budgets. That gap is why institutions are being outpaced, and it’s exposing them to faster, more sophisticated attacks.

How Sardine’s AI risk platform changes the equation

The Sardine platform was built to address these exact weaknesses. We do this by unifying fraud and AML in a single platform, leveraging AI to move faster than fraudsters, and creating a shared fabric of intelligence across the institution.

1. Unified Fraud and AML

Fraud and money laundering are two sides of the same coin. Sardine combines them into a single platform, connecting fraud signals (suspicious deposits, mule accounts, device takeovers) with AML typologies (structuring, layering, suspicious fund flows). This integrated approach ensures institutions can shut down criminal networks instead of just chasing individual alerts.

2. Real-time risk interdiction

Fraud happens in milliseconds, so prevention must too. Sardine enables banks and fintechs to block or hold suspicious transactions instantly, across ACH, wires, RTP, FedNow, cards, and checks. With over 6,000 risk signals, including device intelligence, behavioral biometrics, and anomaly detection, we identify high-risk activity before funds leave the system.

3. Consortium Intelligence at Scale

Fraud exploits the gaps between institutions, moving through mule accounts and across rails faster than any single bank or fintech can track. Our Sonar Consortium closes those gaps by connecting insights from banks, fintechs, processors, and telcos to surface mule activity and repeat attackers across the ecosystem.

That view is strengthened by Sardine’s network of more than 3 billion profiled devices, giving risk teams the context to act before funds disappear. Intelligence is shared in a way that highlights patterns without exposing PII, so institutions can collaborate safely and in line with regulatory expectations.

4. Enterprise-grade AI made accessible

Smaller institutions face the same threats as large banks, but don’t have access to the same engineering resources or data science teams.The Sardine platform levels the playing field with out-of-the-box rulesets and AI models available that can be put to work immediately.

Our AI rule builder enables risk teams to create, modify, and test rules themselves, without support from engineering or vendor professional services. Rules can be written in natural language, deployed in minutes, and improved continuously through feedback loops.

Additionally, our rules and AI models can be tuned to an institution’s unique risk profile and policies, and are delivered in a way that is accessible, transparent, and regulator-ready.

5. Protecting customers without adding friction

Fraud prevention cannot come at the expense of user experience. By analyzing device and behavioral signals invisibly in the background, Sardine reduces false positives and minimizes unnecessary holds, ensuring legitimate customers transact freely while fraudsters are stopped.

Why AI is the only way forward

Fraud evolves faster than static systems can react. Rules hard-code yesterday’s attacks but miss tomorrow’s. Analysts are buried in alerts, models drift, and compliance capacity never scales as quickly as the risks do. Defending a real-time payments system requires defenses that move just as fast. The only way to stay ahead is to fight bad AI with good AI.

AI systems can uniquely help risk teams:

From policy to practice

Regulators are right to call for modernization: risk-based funds availability, confirmation-of-payee services, and standardized fraud data sharing are all important steps. But policy alone will not stop fraud. Technology must operationalize these principles.

That’s where Sardine comes in. Our platform delivers the AI-driven fraud and AML infrastructure that turns regulatory intent into real protection:

• Real-time interdiction to enforce risk-based availability
• Cross-institution intelligence to support a national fraud data framework
• Unified fraud and AML oversight for more effective mule account disruption
• Explainable AI tools that enable safe innovation and adoption by smaller institutions

The regulators’ request for input marks a turning point. The question is no longer whether fraud should be addressed at the infrastructure level, but how quickly we can deploy the tools to make it happen. Sardine is ready to lead that charge.