---
title: "Global bank cuts fraud false positives by 62%"
source: https://iaastha.com/newsroom/fraud-false-positives-62-percent/
type: News
date_published: 2026-08-31
date_modified: 2026-08-31
author: White Sarah
description: "A rebuilt decision layer — not a new model — returned investigator trust and cut alert noise by nearly two thirds."
publisher: iAastha
---

# Global bank cuts fraud false positives by 62%

A top-three global bank has reduced false positives in fraud alerts by 62% after rebuilding its decisioning layer with iAastha.

The programme kept the incumbent detection model and changed everything around it: thresholds tied to investigator capacity, explanations attached to every alert, and a feedback loop that retrains on what investigators actually closed.

A rebuilt decision layer — not a new model — returned investigator trust and cut alert noise by nearly two thirds.

iAastha Press Office

press@iaastha.com

---
Cite as: "Global bank cuts fraud false positives by 62%" — iAastha, https://iaastha.com/newsroom/fraud-false-positives-62-percent/
Site index for AI: https://iaastha.com/llms.txt
