Fraud detection and prevention systems play a crucial role in protecting financial institutions (FIs) from financial losses and reputational damage while ensuring compliance with applicable laws. These tools identify suspicious activity and mitigate fraud to avoid financial damage to the FI or its customers. Effective fraud systems help manage operational costs and support an institution’s growth by enabling more, faster, and lower-friction transactions at a low successful fraud rate.

This Sector Spotlight covers fraud systems for payments rails that FIs directly enable access to for their customers. The three key types of products that are adjacent to transaction fraud management are KYC, AML monitoring, and authentication.

We do not cover fraud systems that are sold primarily to nonbanks, directly or through distributors (e.g., acquiring banks, payment gateways, payment processors) but offer examples in our supplementary list. Many of these systems are designed to mitigate card-not-present fraud (merchants are liable for fraudulent card-not-present transactions by card network rules) and chargeback fraud (the acquiring bank may hold the merchant liable and debit its account). They may also cover ACH transactions in this context.

Bank-focused fraud systems address the following:

• Checks: Check fraud systems offer two core capabilities, which are frequently part of the same solution. One is broadly defined as check-image verification, which confirms that the check is legitimate, validates signatures and amounts, and checks for alterations. The other uses data from the bank and check-fraud consortia to risk-score the check based on historical customer activity and consortium data to screen for the risk of check kiting and payment scams like excess of purchase price, fake payments for sold goods, or lottery scams.
  
  
• ACH: ACH fraud systems block fraudulent or fraud-induced payments before settlement. These systems aren’t necessarily sophisticated: An ACH fraud system may only process a batch file and deny or flag transactions for review based on basic transaction parameters (like transaction amount or type of recipient). A more sophisticated system risk-scores ACH transactions based on historical customer transaction patterns and consortium data, and these systems use some combination of device, location, behavior intelligence, and step-up authentication to prevent or limit damage from fraud.
  
  
• Wires: Wire transfer fraud systems focus on identifying and verifying the payee. The value of the average wire makes the potential damage of authorized push-payment fraud to the customer immense, and the FI faces liability if it’s negligent or careless in processing a wire. Institutions with high volumes of wires, and wires that can be initiated through digital banking, require account takeover and authorized push payment protections like those for ACH, but more robust, and should be tuned to an institution’s wire initiation process and risk tolerance.
  
  
• Real-time payments: Real-time payment fraud systems must be able to immediately process everypush-payment transaction straight-through within the FI’s risk tolerance, protected by robust account takeover prevention and multiple layers of authentication.
  
  
• Cards: Credit and debit card fraud systems use customer behavior and transaction location to help banks and credit unions prevent unauthorized transactions and mitigate chargeback fraud, limiting the bank’s exposure to losses based on network rules and federal regulations. Zero-liability protection for consumer credit cards places substantial liability on issuers for unauthorized transactions.
  
  

What’s going on in transaction fraud systems

As FIs adopt new rails and as fraud technology evolves, so does fraud. Each rail brings different transaction speeds, authentication standards, and processing flows, forcing fraud systems to adapt. Early fraud tools were designed for slower, manual environments like checks and wires, but the growth of digital payments and higher transaction volumes has made automated, real-time detection essential.

Below is a look at the latest developments in fraud systems across categories:

Checks

Check fraud prevention systems have evolved in the two decades following Check 21’s implementation by Regulation CC. By enabling the use of substitute checks and electronic image clearing, Check 21 shifted check fraud management from physical document review to electronic verification. Modern check fraud systems incorporate image analysis tools, based on increasingly sophisticated machine learning algorithms, that flag forged, altered, or counterfeit checks, incorporate check fraud data from consortia, and reference a depositor’s patterns of behavior while taking inputs from different deposit channels.

ACH

The ACH networks process by far the most transactions of any rail. FedACH processed 20.1 billion items in 2024, and The Clearing House’s EPN processed 20.7 billion transactions. ACH fraud isn’t as pervasive as check fraud, however: Depository institutions filed 204,927 suspicious activity reports (SARs) for ACH in 2024, up from 143,269 in 2020. They reported 521,036 check-related SARs in 2024, up from 216,963 in 2020, on lower volume. But faster payments leave less time to claw back a transaction: NACHA estimates that about 80% of ACH network volume settles in one day or less. More sophisticated account takeover protections and machine learning-driven payment fraud detection have made ACH fraud more difficult despite the high transaction volumes.

Real-time payments

Batch-based fraud systems and manual reviews characteristic of check, ACH, and wire fraud management are not sufficient for real-time payments. Key factors influencing fraud systems designed for real-time payments are, first, which institutions originate real-time payments and their risk tolerance, and the overall modernization of transaction fraud software. Low transaction volume relative to other rails creates a low base for fraud (FedNow processed 1.51 million payments in 2024, totaling $38 billion, and RTP processed 343 million payments in 2024, totaling $246 billion). The high volume of real-time payments originated by the largest FIs (TCH-owner banks among them) lends itself to custom fraud solutions, but as other institutions warm to sending real-time payments, third-party solutions should become more popular.

Wires

For FIs that don’t process many wires, there hasn’t been a strong need for wire transfer fraud systems to evolve. Not much on the backend has changed since the mid-1990s, and the FedWire and CHIPS (TCH) networks are still characterized by few transactions, high average transaction value, and intensive manual fraud checks. (In 2024, FedWire initiated 209.9 million transfers with a $5.4 million average value, and TCH’s CHIPS processed 142.0 million total transactions with a $3.3 million average value). But wires are susceptible to push-payment fraud like other electronic payment channels, and modern fraud prevention systems add layers of authentication and integrate sophisticated identity tools to verify the payor and payee.

Cards

According to network rules, liability for potential losses related to card fraud differs for issuing banks, acquiring banks, and merchants, which then dictates the fraud systems they need. Changes to liability, card security, and use of payment channels complicate different players’ needs. (This Sector Spotlight focuses on solutions that are most appropriate for banks involved in card transactions.) Issuers’ adoption of EMV chips in the last decade has lowered the risk of card-present fraud, and merchants have generally been liable for card-not-present fraud. But systems must still detect unusual card-present behavior, and banks need to meet requirements for the 3D Secure authentication protocol.

Transaction fraud systems vendor snapshot

Fraud solutions are typically designed around the types of fraud they handle. Companies frequently offer fraud solutions for more than one payment rail, and features may be offered as modules for an enterprise fraud solution. A bank or credit union’s choice of payment fraud systems depends on its business and functional requirements; payment fraud prevention needs are typically based on rails offered and payment volume.

Here’s a snapshot of transaction fraud systems. The list is representative, and our analysis is based on the best available public information: