An empirical investigation into the impact of artificial intelligence on team competency and collaborative practices across the software development life cycle

Sammanfattning

This study investigates the impact of artificial intelligence (AI)-assisted tools on team competency and collaborative practices across the Software Development Life Cycle (SDLC). With Large Language Models (LLMs) now deeply embedded in everyday engineering workflows, the way teams make decisions, divide responsibilities, and interact with one another is shifting in meaningful ways. Using a qualitative research approach, this study draws on semi-structured interviews with professionals working across various SDLC roles in European software development teams. The data were analyzed using Reflexive Thematic Analysis (RTA), enabling an in-depth exploration of how AI tools influence daily work practices, skill requirements, and team dynamics. This study set out to answer two core questions: how AI-assisted tools are changing the competencies pro-fessionals need across the SDLC, and in what ways they are reshaping how teams collaborate and com-municate. On both fronts, the findings paint a consistent picture — AI is acting primarily as a productivity enabler rather than a replacement for human expertise. Routine tasks are being automated, but this is push-ing human roles toward verification, oversight, and higher-level decision-making rather than eliminating them. The findings reveal that AI is reshaping software development in two deeply connected ways — how people work and how they work together. Rather than replacing human expertise, AI is taking over routine tasks while pushing professionals toward new responsibilities like supervision, governance, and critical evalua-tion of AI outputs. Skills such as prompt engineering are becoming essential, and trust in AI, while growing, remains conditional — people verify rather than simply accept. At the same time, collaboration is shifting, with AI often becoming the first point of consultation, quietly changing how and when teammates actually talk to each other. These efficiency gains come with real trade-offs though, including the risk of deskilling and fewer meaningful peer interactions. Ultimately, the successful adoption of AI in software development is as much about people, culture, and collaboration as it is about the technology itself.