Utilizando LLMs para automatizar a descoberta de oportunidades de captação de recursos para instituições de pesquisa

Abstract

This work investigates the use of Large Language Models to support institutional processes, focusing on the preparation of funding proposals—a complex task that requires extensive research and analysis. The identified problem lies in the difficulty research departments face in locating and organizing funding opportunities, due to the large volume of information and the lack of automated tools capable of processing different calls and notices. The proposed solution is based on Retrieval-Augmented Generation integrated with the Browser-use agent, enabling the automation of the collection, processing, and semantic organization of these opportunities. The study contributes by demonstrating how LLM-based agents can transform research management practices within educational institutions.