TL;DR
Get home office essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
A PYMNTS Intelligence report found that 59% of surveyed companies using at least three AI fraud tools stopped 90% or more of attempted payment fraud before losing money, compared with 32% of companies without AI defenses. The July survey covered 150 U.S. treasury and finance executives at companies with annual revenue of at least $100 million; its findings show an association, not proof that AI alone caused the difference.
Fifty-nine percent of companies using three or more AI tools to combat payment fraud said they stopped at least 90% of attempted attacks before losing money, according to a PYMNTS Intelligence report published in September. The comparable share among companies without AI defenses was 32%, a 27-percentage-point gap in a survey of 150 U.S. treasury and finance executives.
The report, titled Prevention First: Building a Smarter Defense Against Payments Fraud, was produced with Bottomline and is based on a July survey of executives at U.S. companies with annual revenues of at least $100 million. It examined how businesses protect payments to suppliers. The findings describe respondents’ reported outcomes; they do not establish that AI tools alone caused the differences.
Among companies that had adopted AI for fraud prevention, 77% said detection performed better than previous methods. Eight percent said it performed worse. The report did not provide a comparison of financial losses or describe how respondents measured performance, so these results reflect executives’ assessments rather than a common audited measure.
Use of specific tools was widespread among AI adopters: 83% used real-time risk scoring to assess transactions as they occur, 82% used automated document verification, and 78% scanned incoming messages for signs of AI-generated fraud. The report links stronger prevention with combining tools and payment controls, rather than relying on detection software alone.
Why Layered Payment Controls Matter
The findings matter to finance teams because supplier-payment fraud can succeed when criminals exploit a change in account details or persuade an employee to authorize an unusual transfer. Earlier detection may give a company a chance to verify a request before money leaves its account. The reported gap between firms using multiple AI tools and those without AI defenses suggests that some businesses see value in adding automated checks to payment workflows.
But the survey does not show that adding AI by itself will prevent fraud at the same rate for other companies. The report emphasizes that supplier information, account-change verification and approval procedures remain part of the defense. A tool that flags suspicious activity may have limited effect if staff can bypass checks or if supplier records are not reliable. The practical issue for businesses is how technology fits into existing controls and how well those controls are followed.
The findings also point to a substantial adoption gap: 57% of surveyed companies had no AI fraud detection tools. The report says 58% of that group were implementing the technology or expected to adopt it within 12 months. These figures suggest interest among non-adopters, but do not establish whether those plans were funded, completed or effective.
As an affiliate, we earn on qualifying purchases.
Survey Scope and Report Findings
The research focused on business payments to suppliers, not consumer fraud or every form of corporate financial crime. Its sample comprised 150 treasury and finance executives at U.S. firms earning at least $100 million a year. The published material does not describe the survey’s sampling method or say that its results represent all U.S. businesses, so readers should treat the percentages as findings from this surveyed group.
The report gave two examples of defenses working together. In one case, monitoring software flagged an attempted wire transfer after an unexpected account change. In another, a business stopped an effort to redirect an ACH payment because the transaction differed from the supplier’s normal activity. The examples illustrate the process described by the report, but it did not identify the companies or provide independent documentation of the incidents.
Companies also reported planned changes to their controls. 81% planned to improve supplier-onboarding verification, 75% expected to invest in bank-account validation before payment, and 73% planned to enhance AI-driven transaction monitoring. These are stated intentions, not evidence that the measures have since been adopted.
“59% of companies using three or more AI tools said they stopped at least 90% of attempted fraud before suffering a loss.”
— PYMNTS Intelligence report, “Prevention First: Building a Smarter Defense Against Payments Fraud”
real-time transaction risk scoring tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Limits of the Survey Evidence
The available findings do not establish whether firms using multiple AI tools had lower fraud losses because of those tools, or whether other differences—such as staff training, transaction volumes, approval rules or prior investment in controls—also contributed. The report compares survey groups but does not provide enough detail to isolate cause and effect.
It is also unclear how the survey defined an attempted attack, a successful prevention or a loss, and whether responses were checked against transaction records. The published material does not give a margin of error or the proportion of respondents in each group. The examples of stopped payments are not accompanied by company names or independent verification.
Finally, planned adoption and control upgrades are not the same as completed changes. The report does not say which organizations followed through, how much they spent, or whether the investments reduced fraud afterward. The findings should not be read as a guarantee that AI tools will prevent a particular company’s losses.
automated document verification system
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Tracking Adoption and Outcomes
The report identifies intended investments, but the next useful evidence would show whether companies carried them out and whether their payment outcomes changed. Follow-up data could compare fraud attempts, prevented payments and realized losses over a defined period, while accounting for differences in company size and controls. No such follow-up results or publication date were included in the material provided.
For now, companies assessing payment defenses can use the findings as a report on surveyed firms, not as a forecast of their own results. The stated priorities—supplier verification, account validation and transaction monitoring—remain plans for many respondents, and the extent of implementation is unknown. Any assessment of the report’s impact will depend on later evidence about adoption and actual fraud outcomes.
As an affiliate, we earn on qualifying purchases.
Key Questions
What does the 59% figure measure?
It is the share of surveyed companies using three or more AI fraud tools that reported stopping at least 90% of attempted attacks before suffering a loss. It is a survey result, not a guarantee for other firms.
How did companies without AI defenses compare?
Thirty-two percent of companies without AI defenses reported stopping at least 90% of attempted fraud before a loss. The difference from the 59% figure is 27 percentage points.
What tools did AI adopters report using?
Among AI adopters, 83% used real-time risk scoring, 82% used automated document verification, and 78% scanned incoming messages for signs of AI-generated fraud, according to the report.
Does the report prove AI caused better fraud prevention?
No. The findings show an association between using multiple AI tools and reported prevention outcomes. The survey does not establish that AI alone caused the difference or guarantee similar results at other companies.
What is not yet known about adoption plans?
The report says some non-users expected to adopt AI and many firms planned upgrades to supplier checks, account validation and monitoring. It does not show how many completed those plans or whether they reduced losses.
Source: rss
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
