AI in software product management
Nasir, Irzam Yazdan (2025)
Diplomityö
2025
School of Engineering Science, Tietotekniikka
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe20251204114698
https://urn.fi/URN:NBN:fi-fe20251204114698
Tiivistelmä
Artificial Intelligence (AI) is influencing the way software products are planned, developed and evaluated. Traditional software product management (SPM) practices are largely based on the experience of the product managers, input from the stakeholders, and manual analysis. However, the increase in the availability of behavioural data, telemetry and advanced analytics has allowed AI to augment decision-making across the product lifecycle. This thesis investigates the role of AI in contributing to strategic and operational activities in SPM, and also identifies the opportunities, challenges, and risks involved in its adoption.
The research is conducted using a Multivocal Literature Review (MLR), which is a synthesis of peer-reviewed research and high-quality industry publications published from 2016-2025. The review shows that AI supports the management of products through a better prioritisation of the backlog, easier sprint planning, analysing customer insights, and predicting the roadmap. Machine learning models, predictive analytics, and new generative artificial intelligence (AI) tools can help improve the accuracy of forecasts, eliminate repetitive work, and enable teams to react more rapidly to changes in the market.
Despite all these benefits, the literature also points to limitations of data quality, organisational readiness, explainability, ethical issues, and complex integration. Without strong governance and a clear strategy between what AI can do and what the product needs to do, adoption of AI can be fragmented or unreliable. The findings emphasise the fact that AI should work as a decision-support mechanism, rather than as replacement of human judgement.
The thesis contributes by pulling together what is currently known, describes responsible practices for integration and gives suggestions that stress the importance of governance, collaboration among humans and AI, the quality of data, and constant evaluation. These insights are intended to help organisations that are looking to leverage AI to enhance product decision-making, agility and customer value in a sustainable, ethical way.
The research is conducted using a Multivocal Literature Review (MLR), which is a synthesis of peer-reviewed research and high-quality industry publications published from 2016-2025. The review shows that AI supports the management of products through a better prioritisation of the backlog, easier sprint planning, analysing customer insights, and predicting the roadmap. Machine learning models, predictive analytics, and new generative artificial intelligence (AI) tools can help improve the accuracy of forecasts, eliminate repetitive work, and enable teams to react more rapidly to changes in the market.
Despite all these benefits, the literature also points to limitations of data quality, organisational readiness, explainability, ethical issues, and complex integration. Without strong governance and a clear strategy between what AI can do and what the product needs to do, adoption of AI can be fragmented or unreliable. The findings emphasise the fact that AI should work as a decision-support mechanism, rather than as replacement of human judgement.
The thesis contributes by pulling together what is currently known, describes responsible practices for integration and gives suggestions that stress the importance of governance, collaboration among humans and AI, the quality of data, and constant evaluation. These insights are intended to help organisations that are looking to leverage AI to enhance product decision-making, agility and customer value in a sustainable, ethical way.