Application of Bpm and Process Automation to Reduce Fraud Rates in an Insurance Company

Olcese A.V.; Ayras H.O.; Avalos-Ortecho E.; León-Chávarri C. · Advances in Transdisciplinary Engineering (2026) · DOI: 10.3233/atde260757

Resumen

This study addressed the problem of high undetected fraud levels in an Argentine insurance company, which generates annual losses of ARS 1,434 million, by applying Business Process Management (BPM) and process automation to the partial-damage automobile claims workflow. The research focused on redesigning the initial analysis phase, where inconsistencies and manual errors originated. The AS-IS process was mapped and simulated, revealing long lead times, fragmented decision paths, and analyst overload. Based on these findings, a digital Poka-Yoke approach was introduced through blocking smart forms, together with an automated validation engine integrating OCR, rule-based scoring, and automatic document verification. The TO-BE model was validated through Arena simulation, showing a 60% increase in fraud detection (from 4.68% to 7.49%), an 81% reduction in the resolution time of non-fraudulent cases, and a 6.08% reduction in the company's loss ratio derived from annual savings. Several manual activities that did not require analytical judgment were automated, including coverage verification, policy validity checks, and requests for additional documentation, previously identified as a major bottleneck. The redesigned workflow ensures completeness through blocking forms and incorporates targeted fraud alerts that enable earlier identification of suspicious cases, contributing to a more efficient and fraud-aware claims pathway.

Instituciones: Universidad de Lima