Generative AI-assisted software development teams : opportunities, challenges, and best practices
Noor, Nouman (2025)
Diplomityö
2025
School of Engineering Science, Tietotekniikka
Kaikki oikeudet pidätetään.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe2025053056086
https://urn.fi/URN:NBN:fi-fe2025053056086
Tiivistelmä
This thesis investigates the integration of generative artificial intelligence (AI) tools into software development teams, with a specific focus on startup environments. It examines how AI-powered tools like GitHub Copilot and ChatGPT affect team productivity, workflows, skill development, and risk management. The study follows a qualitative research approach using semi-structured interviews with eight professionals working in AI-assisted development roles.
The findings indicate that generative AI can enhance productivity, automate routine tasks, assist in debugging and documentation, and support continuous learning. However, challenges include AI hallucinations, tool-context mismatch, data privacy risks, and skill dependency. Thematic analysis reveals both technical and organizational considerations, and the thesis proposes best practice guidelines for startups: starting with low-risk tasks, maintaining human oversight, developing prompt literacy, and defining AI usage boundaries.
This study contributes actionable insights for teams seeking to integrate AI into their development lifecycle while preserving software quality, team autonomy, and long-term innovation capabilities.
The findings indicate that generative AI can enhance productivity, automate routine tasks, assist in debugging and documentation, and support continuous learning. However, challenges include AI hallucinations, tool-context mismatch, data privacy risks, and skill dependency. Thematic analysis reveals both technical and organizational considerations, and the thesis proposes best practice guidelines for startups: starting with low-risk tasks, maintaining human oversight, developing prompt literacy, and defining AI usage boundaries.
This study contributes actionable insights for teams seeking to integrate AI into their development lifecycle while preserving software quality, team autonomy, and long-term innovation capabilities.