Aplicação de Ferramentas Baseadas em Inteligência Artificial para Minimização de Desafios e Otimização de Processos na Engenharia de Software

Abstract

Software Engineering faces significant challenges in the initial development phases, especially in requirements definition and prototyping, which frequently results in rework, inconsistencies, and increased costs. Given this scenario, this work aims to analyze the difficulties present in these stages and propose solutions based on Artificial Intelligence tools, highlighting NotebookLM and Gemini Canvas. The research adopted an applied approach, beginning with a theoretical investigation into requirements, prototyping, and AI, followed by an analysis of the tools' functionalities, and finally, the simulation of a fictitious scenario of a Condominium Management System. Structured files were created to serve as a knowledge base, enabling the application of prompt engineering techniques for the validation, organization, and generation of requirements, as well as the creation of functional prototypes. The results demonstrated that NotebookLM contributes to reducing ambiguities and streamlining requirements analysis, while Gemini Canvas facilitates system prototyping. The study concludes that the use of these tools can minimize common failures in the initial stages of software development, making the process more efficient, precise, and aligned with the project's needs.