AI Simulations for Transaction Reversals Based on Trust

ChangeBridge: Patent Apps - Business Methods (G06Q)

Detected March 26th, 2026

Summary

The USPTO has published a patent application detailing a system and method for implementing transaction reversals using AI simulations based on trust levels between users and recipient accounts. The application, filed on September 24, 2024, aims to validate or deny transactions based on AI-driven trust assessments.

View original document View source feed page

What changed

This document is a patent application (US20260087498A1) from the USPTO describing a novel method for transaction reversals using AI. The system assesses trust levels between a user and recipient accounts to simulate and validate transactions before forwarding them to a payment gateway. If trust thresholds are not met, the transaction is denied, and alerts are sent.

While this is a patent application and not a regulation, it signals potential future technological developments in financial transaction processing. Compliance officers in financial services should be aware of emerging AI applications that could impact transaction security, fraud prevention, and customer trust mechanisms. No immediate compliance actions are required, but monitoring such technological advancements is prudent for understanding future regulatory landscapes and operational risks.

Source document (simplified)

← USPTO Patent Applications

SYSTEM AND METHOD FOR IMPLEMENTING TRANSACTION REVERSALS THROUGH ARTIFICIAL (AI) SIMULATIONS BASED ON SET THEORETIC PROPERTIES

Application US20260087498A1 Kind: A1 Mar 26, 2026

Inventors

Maharaj Mukherjee, George Albero, Naga Vamsi Krishna Akkapeddi, Syed Najumudeen

Abstract

Implementing transaction reversals through (AI) simulations based on set theoretic properties is provided. The method receives a user request including user identification of a plurality of target recipient accounts. A discrete level of trust exists between the user and each of the target recipient accounts. The method further includes receiving a user specification of a transaction. A target recipient account is auto-selected based on the level of trust. The method derives identification information from the user request. The identification information includes user identity, recipient identity, and recipient know-your-customer information. The method leverages AI to perform a simulation of the transaction. When the simulation obtains greater than a threshold of trust between the user and the recipient, the method validates the transaction and forwards the transaction to a payment gateway. When the service is not confirmed, the method reroutes the transaction from the payment gateway for transaction denial, and sends alerts.