Fraud–AML Convergence (FRAML)

In most institutions, fraud and anti-money-laundering grew up as separate teams with separate tools and separate goals. Fraud protects the institution and its customers from losses; AML protects the financial system from illicit money and reports it to the government. In fintech, that wall makes less and less sense, because both teams look at the same customers, the same transactions, and the same signals, just with different questions.

The industry shorthand for merging them is FRAML, fraud plus AML. This lecture is about why convergence matters, where the two disciplines connect, and the AML obligations that fraud can quietly trigger, which fintechs sometimes miss because the fraud team and the AML team aren't talking.

Why does fintech specifically drive convergence? First, shared signals: the device fingerprint, the velocity, the identity anomalies, the network connections that flag fraud are the same data that flag laundering, so analyzing them twice in two silos is wasteful. Second, speed: instant, irreversible payments force both fraud and AML decisions to happen in real time at the moment of the transaction, not in a next-day batch, which pushes the controls into the same decisioning layer.

Third, economics: a lean fintech can't afford two fully separate stacks and teams duplicating effort. And fourth, the threat itself: the same actor running an account-takeover or a scam is often also moving the proceeds, so the fraud event and the laundering event are two views of one criminal scheme. Converging the functions lets you see the whole picture instead of two halves.

Here's the critical compliance point fintechs miss. Fraud isn't only a loss to absorb; it generates dirty money and reporting duties. The proceeds of fraud are exactly the illicit funds the BSA exists to track, and moving them is money laundering.

So scams, account takeover, business email compromise, romance and investment fraud, these aren't just fraud tickets; when you detect them, they frequently require a suspicious activity report under the SAR rules, not merely a refund or a chargeback. Money mules, accounts opened or rented to receive and forward proceeds, are a textbook AML typology that lives right at the fraud-AML seam. FinCEN has issued advisories on many of these patterns to guide reporting.

The failure mode is a fraud team that quietly closes an account for 'fraud,' recovers what it can, and never tells AML, so a reportable event goes unreported. That gap is a real BSA violation.

Let's name the failures. The biggest is the silo: fraud and AML in separate systems with no shared escalation path, so neither sees the other's view of the same customer. From that flows the classic error, closing or remediating a fraudulent account purely as a fraud matter and never evaluating it for a SAR, which is exactly the reporting gap regulators look for.

Another is optimizing narrowly for fraud loss, reimbursing a victim and moving on, while ignoring that the money flowed through a network that's also laundering. And mule detection in particular suffers in silos, because catching mules requires connecting accounts and flows across the platform, which a single-account fraud view never does. Converged FRAML teams, by contrast, can see one actor across fraud and laundering and route the same event to both a recovery action and a SAR decision.

So how do you do FRAML well? Start with a shared data and signal layer, so fraud and AML draw on the same device, identity, velocity, and network intelligence rather than rebuilding it twice. Build a joint escalation path: every confirmed or suspected fraud case automatically gets a suspicious-activity review, so nothing is closed as 'just fraud' without an AML decision.

Maintain a mapping of fraud typologies, account takeover, scams, business email compromise, mule activity, to the AML reporting obligations they trigger, so analysts know when a fraud finding becomes a SAR. And here's the nuance: converge the tooling and the teams, but keep the distinct legal duties clear. Fraud and AML have different goals and different rules; FRAML is about a shared lens and shared data, not about letting the SAR obligation get lost inside a fraud workflow.

Done right, convergence makes you both more efficient and more compliant.

Let's lock it in. Fraud and AML look at the same customers and signals, and fintech's speed, shared data, and lean teams push them to converge into FRAML. The compliance heart of it is that fraud generates dirty money, so a detected fraud frequently triggers a suspicious activity report, not just a chargeback, and money mules sit right at the seam where silos fail.

The fix is a shared data and signal layer plus a joint escalation path that sends every fraud case through a SAR review, while keeping the distinct legal duties explicit. Self-check: when your fraud team shuts down a scam account, does an AML analyst automatically evaluate it for a SAR? If not, that's a reporting gap to close today.

Next, we make the reporting itself concrete: SAR and CTR obligations in a fintech.