INTELLIGENT API WITH GENERATIVE AI AND RAG FOR QUERYING BUSINESS DOCUMENTS

Abstract

This work presents the development of an API capable of processing internal corporate documents in PDF and Excel formats, making them searchable and accessible through intelligent queries with Generative Artificial Intelligence. The proposed architecture applies the RAG (Retrieval-Augmented Generation) methodology, integrating techniques of text extraction, optical character recognition (OCR), database storage, embeddings generation, and orchestration mechanisms for intelligent queries. As a result, corporate documents become knowledge bases for contextual responses, overcoming the limitations of traditional semantic search and ensuring greater accuracy and speed for the user. The proposal was validated through a case developed in an academic environment, demonstrating potential for adaptation in companies from different sectors