IMPLEMENTATION OF ARTIFICIAL INTELLIGENCE IN THE CLASSIFICATION OF PRODUCTS IN NF-E BY SEFAZ/CE: A TECHNICAL REPORT ON INNOVATION AND COMBATING TAX EVASION

The Secretariat of Finance of Ceará (SEFAZ/CE) implemented an Artificial Intelligence (AI) solution to improve the classification of products in Electronic Invoices (NF-e), aiming to strengthen collection and combat tax evasion. The project sought to improve the collection, inspection and taxation processes, ensuring accurate identification of products and scalability in the face of large data. The automatic sorting solution sped up the process and ensured greater accuracy while reducing human intervention. The results include an average accuracy of 90.18% in the classification of products, the basis for updating reference values for items such as beer and soft drinks, and an aid to tax actions that resulted in significant collection. The impact of the project promotes greater tax justice and transparency in public administration, contributing to the understanding of how AI can be incorporated into public administration to address contemporary challenges.