AI Made B2B Fraud Cheap And Trust Expensive

TL;DR

Prime Big Deal Days · Oct 6–7Offer from Amazon

Get smart everyday buys delivered free — and shop member deals

• Fast, free delivery on millions of items
• Access to Prime Big Deal Days deals on October 6–7
• Prime Video, Amazon Music and more included

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

A PYMNTS interview with Bottomline risk and fraud officer Katie Elliott describes how AI can make business payment fraud attempts broader and faster, while manual checks remain costly to scale. Elliott argues that layered supplier verification and targeted human review may help finance teams manage the risk, but the report provides no data measuring fraud rates or the effectiveness of these approaches.

AI is lowering the cost of scaling fraud attempts, while manual checks remain expensive for finance teams to expand, Bottomline senior risk and fraud officer Katie Elliott said in an interview published by PYMNTS. The report argues that businesses may need to rely more on layered supplier verification and reserve staff review for unusual payment activity, particularly as faster payments can leave less time to recover funds.

Elliott said criminals are using AI and other available tools to increase the reach and speed of their attempts. She described a shift from more targeted schemes toward mass phishing and spamming. The interview does not quantify how often AI is being used in B2B fraud or how much attack volume has changed; its account is based on Elliott’s assessment of the threat.

The report focuses on a mismatch in operating costs: attackers can send many attempts and need only some to succeed, while a company must check legitimate payments without holding them up or building a large fraud-review team. Elliott said that once a payment is authorized, it may move quickly, and a recipient can withdraw or move the funds soon after. That makes pre-payment checks more valuable than relying on recovery after a payment has gone to the wrong party.

For those checks, Elliott recommended using multiple types of information rather than relying on a single bank-account validation. Her examples included digital identity, phone details and email-domain history, alongside other signals that can help judge whether a payment request fits the supplier relationship. She also said third-party data and verification services can be costly, especially for smaller businesses, and suggested payment networks could spread those costs across more transactions.

Why Supplier Checks May Move Upstream

The issue affects how companies approve payments, not only how they investigate fraud after a loss. If a payment can move quickly and be difficult to recover, finance teams have a stronger reason to establish confidence in a recipient and payment instruction before money is sent. That can place more pressure on accounts-payable processes to detect altered bank details, unexpected requests and other departures from normal supplier behavior.

There is also a cost question for CFOs. Building a system that combines identity, account and behavioral information may be difficult for one business to fund and maintain. Elliott’s argument is that shared verification services offered through payment networks could make some tools more accessible. The interview presents this as a potential role for networks, not evidence that shared systems already prevent fraud at a measured rate or will be less expensive for every company.

The proposed approach also changes how staff time is used. Instead of manually checking every payment to the same degree, automated systems could process routine activity and direct unusual cases to people. That could focus employees on the situations where context and judgment matter, though the source does not establish how accurately AI systems can distinguish normal activity from fraud or how much human review is needed.

From Payment Recovery to Prevention

The report describes a change in emphasis within accounts-payable fraud controls. Traditional procedures can include checking supplier details during onboarding and validating payment instructions. Elliott’s concern is that convincing impersonation and faster payment movement may weaken the value of checks performed only at the start of a supplier relationship. She said trust may need to be re-established over time, especially when payment details or behavior change.

Her comments came during a discussion for the PYMNTS B2B Payments Event 2026, titled “Stopping B2B Fraud Before AI Moves the Money.” The source is an interview and industry analysis, not a published fraud study: it provides no sample, comparative data, named incidents or independent measurement of AI-enabled attack growth. Its specific recommendations and descriptions of the threat should be understood as Elliott’s views.

“They are using the technology that’s out there, the AI, every tool available to them in order to make their attempts bigger, broader, faster.”

— Katie Elliott, senior risk and fraud officer at Bottomline, speaking to PYMNTS

Fraud Scale and Tool Performance

The interview does not say how widespread AI-enabled B2B fraud is, how quickly it is growing, or which types of businesses face the greatest exposure. It gives no figures for losses, successful attacks, recovery rates or the cost of particular verification services. The account describes a risk and a proposed response, rather than documenting a measured change across the market.

It is also unclear which payment networks currently offer the proposed combination of verification tools, what those services cost, and how much they reduce fraud without delaying valid payments. Elliott’s example of routing an unusually large request to a person illustrates an approach, but the source does not report tests of its accuracy or explain how a company should set thresholds. The effectiveness of automated review and the extent of human oversight required remain open questions.

How Finance Teams Could Respond

The immediate steps described by Elliott are to avoid basing payment decisions on one data point, track changes in supplier instructions and transaction behavior, and send meaningful anomalies for human review before authorization. Companies considering shared tools would need to compare their costs and capabilities with the resources required to operate verification independently.

The source identifies no new regulation, product launch, deadline or follow-up study. Further evidence would be needed to show whether network-based verification can lower costs for smaller businesses and whether automated systems can catch suspicious changes while allowing routine payments to proceed. For now, the report’s central message is a warning and an operational proposal—not proof that a particular model has solved the problem.

Key Questions

What is the development in this report?

A PYMNTS interview with Bottomline’s Katie Elliott describes how AI may let fraudsters scale attempts more cheaply and argues that businesses should strengthen supplier verification and focus human review on anomalies.

Does the report prove that AI-driven B2B fraud is increasing?

No. Elliott says attackers can make attempts “bigger, broader, faster,” but the report provides no market-wide data on attack volume, losses or growth rates.

What checks does Elliott recommend?

She advises using multiple signals, including digital identity, phone information and email-domain history, rather than relying on a single account detail. Unusual payment changes or activity could be directed to staff for review.

Why does payment speed matter?

Elliott says authorized payments can move quickly, limiting the time available to stop or recover funds. Her argument is that checking recipients and instructions before sending can become more important.

What remains unknown?

The report does not quantify the scale of AI-enabled fraud or show how well particular verification systems work. It also does not compare the costs or results of network-based tools with systems built by individual companies.

Source: rss