Early Detection: Why Top-Performing Firms Focus on Fraud Before It Starts

It’s never been harder for businesses to manage the integrity of their payment operations. Fraud schemes are evolving as bad actors leverage artificial intelligence, making threats more difficult to detect in real time. At the same time, faster payments leave businesses with less room for error and are more reliant on automation—one of the very forces fueling AI-driven fraud. Only 8% of surveyed companies report no problems on this front.

As a result, pressure is building in the payments industry for stronger “upstream” safeguards that can detect identity and bank account risks early in the transaction process. The urgency arises amid warnings from regulators, including the Financial Crimes Enforcement Network, about the growing use of deepfakes and synthetic identities to bypass financial institutions’ verification systems.

It’s why the National Automated Clearing House Association, which governs, develops, and administers the rules for the ACH Network, put new rules into effect in March that require all non-consumer ACH participants to strengthen their monitoring of fraudulent transactions. The rules require two new standardized payment labels—payroll or purchase—on ACH transactions to help banks spot anomalies. Aimed at combating “credit-push” fraud, where criminals trick businesses into initiating seemingly legitimate ACH payments, participants must establish risk-based monitoring processes across all transaction types, with the rule phased in on March 20, 2026, for large originators and on June 22, 2026, for everyone else.

All of this puts greater focus on the steps that verify the process for transferring money to a business or customer for goods or services delivered. The AR function is a delicate balancing act involving speed, customer experience, and risk management. Most companies still rely mainly on detecting fraud or nonclearance after settlement, when losses have already occurred, and recovery options are limited and uncertain. As a result, operational friction is rising, and the cost of missed or delayed detection is becoming more visible.

PYMNTS Intelligence’s latest Certainty Project research shows that leading middle-market firms with at least $100 million in annual revenue are shifting their focus much earlier in the transaction process. By verifying bank accounts and identities earlier—even before payments are initiated or authorized—they can block AR integrity risks rather than absorb losses and attempt recovery later.

These are just some of the findings detailed in “Early Detection: Why Top Performing Firms Focus on Fraud Before it Starts, sponsored by Plaid. This edition examines the challenges surrounding fraud prevention and identity verification in AR processes. It draws on insights from a survey of 60 heads of payments at U.S.-based middle-market companies with annual revenues between $100 million and $1 billion. The study was conducted from March 18-30, 2026.

• Stuck in Reactive Mode
• The Verification Imperative
• Uncertainty Amplifies the Pain
• Read More
• Methodology

Stuck in Reactive Mode

For most firms, AR integrity remains a backward-looking process: Nearly six in 10 businesses mainly detect fraud or nonclearance after settlement.

When a business sends out goods or services and waits to receive payment, there’s a window of time during which things can go wrong. More than half of the businesses surveyed said they typically discover a problem only after the payment has already settled, which can mean the money is gone and getting it back is an uphill battle.

What separates the companies that catch problems early from those that don’t is how well they verify who they’re dealing with. The firms with better early-warning systems aren’t just doing more legwork—they’re also using technology that connects directly to a customer’s bank to confirm, in real time, that the account exists and belongs to the right person. This kind of check, combined with automated fraud scoring, gives companies a chance to pause or redirect a risky payment before it’s too late. Firms that rely mainly on manual document reviews and phone confirmations, while common, tend to be the ones still learning about problems after the fact.

For nearly six in 10 (57%) firms surveyed, fraud or nonclearance is typically detected after respondents treat the payment as “done” and only learn of the failure later via returns, disputes, or outreach from the receiving bank. Only three in 10 usually find these issues early, either before the transaction is initiated (17%) or during authorization (13%). High-uncertainty firms are slightly more likely than others to detect fraud or nonclearance early, but late detection is the norm across all three certainty tiers.¹ Results are also very similar across revenue brackets.

AR integrity challenges extend well beyond post-settlement fraud detection.

More broadly, integrity challenges affect AR for nearly all the businesses surveyed, with 88% reporting at least one issue in the last 12 months. Account quality and clearing failures represent the most frequently encountered problems. Seven in 10 firms experienced either ACH returns due to invalid or closed accounts, or to customer input errors that prevented clearing. Nearly as common are after-delivery reversals and disputes, at 63%.

The data also reveals that about half of the businesses surveyed (47%) reported issues with account identity and legitimacy. Challenges in this area rise markedly with uncertainty levels, something not consistently seen in the other areas. For example, only 8% of low-certainty firms report detecting payments in the last 12 months made with fraudulent or stolen bank accounts. This rate more than doubles for medium- and high-uncertainty firms.

The Verification Imperative

Early detectors are about twice as likely to use bank account ownership verification methods as firms in the post-settlement detection.

There’s no doubt that middle-market firms have at least some systems in place to verify incoming payments. But since so many businesses detect AR integrity threats after settlement, it’s obvious their current methods don’t always perform well.

Bank-integrated account ownership verification stands out as the strongest differentiator for early detection best practices. Roughly eight in 10 firms that usually detect fraud or nonclearance before settlement use instant bank account verification (81%) and open banking-based ownership verification (76%). In each case, adoption is roughly double that seen among post-settlement detectors, at 47% and 35%, respectively. Micro-deposit verification follows a similar pattern, though adoption is much lower overall: 38% in the pre-settlement group and 24% in the post-settlement group.

Real-time identity verification and automated Know Your Customer (KYC) and Know Your Business (KYB) checks also play important roles in early detection. But the gaps between pre- and post-settlement detectors are substantially smaller due to broader adoption. Meanwhile, more basic, widely used practices such as manual document review (used by 95% of firms) and direct customer confirmation (88%) are in place at nearly all firms, with no significant difference in adoption between pre- and post-settlement detectors.

Overall, we also note that early detectors use, on average, eight verification methods, compared to six for the post-settlement group. That means the best performers aren’t just using better tools—they have a more robust toolkit for mitigating AR integrity risks.

Firms that adopt upstream verification tools overwhelmingly report strong results.

The positive impact of using integrated bank account verification becomes even clearer when we look at the firms that have recently adopted it. Among middle-market companies that used instant or real-time bank account verification in the last 12 months, 84% rate this as very or extremely effective in reducing fraud risk or improving payment integrity. Following close behind, 82% of the firms that deployed artificial intelligence or machine learning fraud scoring say it is highly effective, as do a similar share that implemented or upgraded identity verification at customer onboarding (78%).

It’s no coincidence that all three of these methods are “upstream” controls that help protect against AR integrity risks long before the downstream stage of settlement. In fact, two of them, bank account verification and identity verification during customer onboarding, happen before a transaction even begins.

Uncertainty Amplifies the Pain

Average costs of payment fraud and nonclearance double for high-uncertainty firms.

Late detection of payment missteps carries tangible costs for businesses. By the time a fraudulent transaction settles, the seller will likely have already delivered the goods or services. That means either taking a loss or diving into a time-consuming dispute with no guarantee of any recovery. Nonclearance results in lost sales and disappointed customers, which impact the bottom line.

On average, the firms surveyed report that payment fraud and nonclearance cost them 31 basis points of revenue. This is a larger number than it might seem at first glance, since revenue reflects raw sales volume rather than profit.

The impact of uncertainty becomes very clear in this context. High-uncertainty firms incur costs equal to 42 basis points of revenue from payment fraud and nonclearance, versus just 21 basis points for low-certainty businesses.

For many firms, a major challenge is that faster payments accelerate the pace of fraud and other risks. On average, businesses that agree that advances in payment speed have increased their exposure to fraud say AR integrity costs them 41 basis points of revenue, around 60% higher than other firms. Another important issue is integration. Businesses say their payment verification and fraud detection tools are not well-integrated into their AR workflows, reporting average costs of 40 basis points of annual revenue, compared with 30 to 35 for the rest. These findings highlight the importance of partnering with identity verification and fraud prevention solution providers to help a firm achieve effective implementation for its specific needs.

Read More

PYMNTS Intelligence is the leading provider of information on the trends driving innovation in finance, banking, and digital payments. To stay up to date, subscribe to our newsletters and read our in-depth reports.

• Waiting for Certainty: Why Most CFOs Are Holding Back on Crypto and Stablecoins
• How Middle-Market Business Uncertainty Rewrote 2025
• Tariffs Turn Up the Heat as Product Leaders Confront Peak Uncertainty

Methodology

“Early Detection: Why Top Performing Firms Focus on Fraud Before It Starts,” the latest installment of the 2026 Certainty Project, is based on a survey of 60 heads of payments conducted from March 18-30, 2026. The survey polled executives at U.S.-based companies with annual revenues between $100 million and $1 billion. The report examines challenges related to fraud and identity verification in accounts receivable processes.

For too long, AR has been treated as a back-office function where fraud and payment failures are something you respond to rather than prevent. A return comes in, a dispute gets filed, the finance team chases a recovery that often never lands. That model worked when ACH was slow enough to give you time to spot trouble. It doesn’t work anymore. Faster payments and AI-driven fraud have collapsed the window, and firms have to update their technology to match.

“The missing piece is up-front detection. The firms catching problems early are verifying bank accounts, matching identities, and assessing fraud risk in real time before the payment settles, and increasingly before it’s even initiated. 57% of firms still detect issues only after settlement, which is a sizable gap and a real opportunity. The ones who’ve closed it are roughly twice as likely to be using instant account verification and open banking-based ownership checks, and 84% of recent adopters call those tools highly effective.

“Up-front detection is a layer that runs between the customer giving you a bank account and you accepting a payment from it. That layer is what we build at Plaid: confirming the account exists and is owned by the right party, scoring the risk in real time, and verifying the identity behind it. None of this is exotic, and none of it requires tearing out existing AR workflows. It just has to be in place before money moves.

“The playbook is on the table. The firms that adopt it earliest will compound the advantage in cleaner receivables, faster settlement, and fewer surprises on their income statements. The rest will spend the next few years catching up.”

^{1. PYMNTS Intelligence defines “uncertainty” as corporate executives’ assessments of unpredictability or lack of assurance in critical business areas, including accounts payable and receivable, cash and liquidity positions, macroeconomic conditions, consumer and customer demand, risk management, compliance and regulatory issues, supply chains, payments capabilities, exchange rates and competitive positions.↩}