Proposta para automatização do processo de leitura de notas fiscais em uma empresa de logística de Manaus.

Abstract

This study proposes an automation solution for the reading and verification of tax invoices in a logistics company located in Manaus, addressing operational challenges caused by technological limitations and the intensive reliance on manual data entry. The problem identified concerns the inefficiency of manual processing, which leads to recurrent rework, transcription errors, and increased processing time due to the inability of current optical readers to interpret digital documents and images with sufficient accuracy. The main objective is to develop a technological approach capable of reducing operational time, improving data reliability, and minimizing human intervention. The methodology is structured as an applied case study with a qualitative approach, using tools such as the Ishikawa Diagram for root-cause identification, process flow mapping for operational analysis, and Monte Carlo Simulation in MATLAB to statistically model three months of operational variability. The results indicate that the automation of document capture and OCR-based data extraction can reduce the average processing time to 2,3 minutes per invoice, minimizing operational bottlenecks and reducing the need for manual rework. The findings demonstrate that the proposed automation presents technical feasibility and has significant potential to enhance operational efficiency, despite not being implemented in practice due to time and resource constraints.