Year: 2025
Type: dissertation
Abstract: This thesis investigates the integration of artificial intelligence (AI) tools, specifically the Cursor AI assistant, in software development processes to address the challenges of accelerating development cycles while maintaining code quality and security.The rapid emergence of large language models (LLMs) has transformed software engineering, yet realworld adoption reveals complex trade-offs between automation benefits and risks such as security vulnerabilities and reduced developer oversight.This study focuses on understanding where and how AI tools can be effectively embedded within development workflows, balancing automation with human expertise.The research was conducted through a multi-project investigation involving four distinct software development projects representing diverse domains: a pharmacy inventory management system utilizing AI-enhanced data processing, an automation system for product entry built on an AI-driven browser control platform, a document management application leveraging AI-assisted optical character recognition (OCR) and a technology company website with planned backend content management and chatbot features.Using an Action Design Research methodology, the study examined the interactions between developers and AI, techno-organizational factors influencing implementation and the relative effectiveness of AI assistance across different development phases.
Author: Gonçalo de Lima Gomes e Boa-Nova
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7112203697