From backlog refinement to ai-augmented decision-making : how generative ai is transforming the product owner role : a case study in large-scale agile software development
Huis in 't Veld, Benjamin (2026)
Diplomityö
2026
School of Engineering Science, Tietotekniikka
Kaikki oikeudet pidätetään.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe2026052857698
https://urn.fi/URN:NBN:fi-fe2026052857698
Tiivistelmä
This thesis examines how generative artificial intelligence (AI) is influencing the Product Owner role in large-scale Agile software development. The study addresses a gap in existing research by focusing not on AI as an autonomous decision-maker, but on how Product Owners use AI in practice to support refinement, prioritisation, coordination, and communication in a complex organisational setting. The research was conducted as a qualitative single-case study in Nokia Mobile Networks and is based on nine semi-structured interviews with Product Owners working in different product contexts.
The findings show that AI is currently used primarily for preparatory and information-intensive tasks, such as internal information retrieval, summarisation, drafting, and early refinement support. Rather than replacing Product Owner judgement, AI mainly supports decision preparation by helping structure fragmented information and reduce manual effort. At the same time, trust in AI is conditional and strongly shaped by validation practices, transparency, task criticality, and the ability to justify AI-assisted outputs in collaborative decision settings. The results further show that responsible adoption depends on socio-technical conditions, including governance constraints, policy-compliant internal tools, workflow integration, and shared organisational practices.
The thesis contributes empirical insight into AI-augmented Product Ownership and shows that the key challenge is not whether AI can generate useful outputs, but how AI support can be embedded responsibly into existing Agile coordination structures while preserving human accountability, transparency, and control.
The findings show that AI is currently used primarily for preparatory and information-intensive tasks, such as internal information retrieval, summarisation, drafting, and early refinement support. Rather than replacing Product Owner judgement, AI mainly supports decision preparation by helping structure fragmented information and reduce manual effort. At the same time, trust in AI is conditional and strongly shaped by validation practices, transparency, task criticality, and the ability to justify AI-assisted outputs in collaborative decision settings. The results further show that responsible adoption depends on socio-technical conditions, including governance constraints, policy-compliant internal tools, workflow integration, and shared organisational practices.
The thesis contributes empirical insight into AI-augmented Product Ownership and shows that the key challenge is not whether AI can generate useful outputs, but how AI support can be embedded responsibly into existing Agile coordination structures while preserving human accountability, transparency, and control.