Uso de ChatGPT no desenvolvimento low-code: um estudo exploratório sobre o uso de IA para desenvolvimento na plataforma Salesforce

Abstract

The digital evolution requires companies to constantly innovate and adopt new technologies to remain competitive in the market. However, one of the main challenges is the lack of skilled professionals, especially in software development, where demand exceeds supply. To mitigate this issue, Low-Code Development Platforms (LCDPs) have gained prominence by enabling the creation of applications with little or no need for traditional programming. Although this approach simplifies development, challenges such as limited customization, difficulties in integrating external systems, and the learning curve for developers still pose barriers to its broader adoption. Given this scenario, this study proposes the integration of Generative Artificial Intelligence (GenAI), specifically ChatGPT, as a support tool for software development in low-code platforms, with a focus on Salesforce. The research adopts an exploratory and experimental approach, analyzing the impact of prompt engineering on development efficiency. The study includes a review of related works, the implementation of a Salesforce application using optimized prompts for code generation, process automation, and debugging, as well as an evaluation of the obtained results. The results indicate that the use of ChatGPT contributed to accelerating development, reducing common errors, and facilitating the understanding of concepts, especially for less experienced developers. Moreover, the interviews revealed positive perceptions regarding the tool’s role as technical and educational support, although the need for human validation in certain contexts is still recognized.