NEW YORK CITY, NY / ACCESS Newswire / November 5, 2026 / Trustmi, a leader in behavioral AI for payment security, today released new findings revealing that $1 in every $48 of enterprise payment value involved an error that Trustmi prevented. Across $240 billion in payments protected over one year, Trustmi prevented $5 billion in payment errors, highlighting how costly mistakes can move through established accounts payable (AP) processes despite existing controls.

The findings show that duplicate invoices and payments accounted for approximately half of affected payment value, followed by data-entry errors (20%), currency mismatches (20%), and payments involving the wrong business entity or missing processing steps (10%).

"Despite long established payment processes, checks and controls, there's a clear pattern of significant payment errors that continue to fly under the radar," said Eli Ben Nun, CTO of Trustmi. "More disparate checks aren't the answer, but errors don't have to be an inevitable cost of business. Using behavioral AI, it's possible to spot fraudulent payments and those made in error, before the money moves."

Where Duplicate Payments Slip Through Existing Controls

To understand how duplicate payments slip through established financial processes, Trustmi retrospectively analyzed one year of ERP records from a global enterprise. The organization had established AP processes and ERP duplicate-detection controls and was already using Trustmi for fraud protection, but not for pre-payment error prevention.

The analysis identified around 85 probable duplicate invoice cases, including 59 cases where both copies cleared a payment run and approximately $560,000 in flagged second-payment value. The findings revealed:

• 61% of cases paid twice had invoice numbers that were not identical. Small changes, like added letters, formatting differences, or revisions made probable duplicates seem like new transactions.
• 58% of duplicates paid twice were entered into the ERP within one week of the original invoice. In many cases, both the original invoice and its probable duplicate were in the system well before payment, often with the same invoice date. More than half cleared the same payment run.
• Eight in 10 duplicates paid twice followed the standard purchase order (PO)-matched invoice process. Matching, approval, and duplicate checks each serve an important purpose, but established procedures did not consistently prevent probable duplicate payments from moving forward.
• Duplicate instances follow predictable seasonality. Duplicates tend to peak around year end, December and January, and again in July.
• Duplicate errors happen most often on a Friday or Monday. Second copies are entered most often on a Friday or a Monday, and least often on a Wednesday.

Payment Error Detection Requires Full Historical Context

The findings highlight the limitations of traditional AP controls. Individual systems can only validate within a record of their own transactions - and without full payment history across an organization, can often miss errors. By connecting records across systems - payment activity, invoice patterns, transaction amounts and vendor details - behavioral AI can help organizations identify probable duplicates and other potential errors.

Trustmi's findings reinforce that preventing payment errors isn't simply a matter of adding more checks but rather enabling widespread access to a vendor's complete payment history to catch anomalies before the money moves.

To explore the complete findings, visit this link.

About the Analysis

Trustmi analyzed $240 billion in enterprise payment value processed over one year, including $5 billion in payment errors prevented before release. Error categories reflect estimated shares of affected payment value, not incident counts. A separate retrospective investigation examined one year of ERP records across 20 business entities in 15 countries, identifying probable duplicate invoices through similarities in vendor information, payment amounts and invoice numbers while excluding recurring monthly charges. Flagged amounts represent potential duplicate-payment value, not confirmed financial losses, and the investigation did not establish whether payments were subsequently recovered.

About Trustmi

Trustmi is the leading behavioral AI platform for payment security, helping enterprises stop fraud and costly payment errors before money moves. By correlating signals across email, vendor behavior, financial documents and payment workflows, Trustmi detects sophisticated threats and payment anomalies that traditional security and finance controls miss. Trusted by global enterprises with complex payment environments, Trustmi protects over $240 billion in payments annually and has helped prevent more than $1 billion in fraud and $5 billion in payment errors. Founded in 2021, Trustmi is headquartered in New York City and backed by leading investors including Cyberstarts and Insight Partners. For more information, visit https://www.trustmi.ai/.

Media Contact

PAN for Trustmi
trustmi@pancomm.com

SOURCE: Trustmi