AI in Product Management : Enhancing the usability of AI tools

Sammanfattning

This thesis examines how artificial intelligence (AI) systems are adopted and experienced in institutional contexts, focusing on usability. A comparative qualitative case study was conducted across three cases: an institutional AI guide (KTH AI Guide), a literature review prototype, and a CBAM reporting platform. Findings show that usability depends less on technical performance than on upstream product-management decisions, including onboarding design, feedback mechanisms, and organizational alignment. Across cases, recurring trade-offs emerged between transparency and complexity, adaptability and governance, and speed of delivery and long-term trust. The study contributes to theory by introducing a Usability-Centered AI Product Management Framework that integrates four core usability dimensions, transparency, learnability, adaptability, and interface complexity, with product-management levers such as scoping, governance, and iterative development. Practically, the framework provides strategies for embedding usability into AI product lifecycles, helping product managers align system functionality with user needs and institutional requirements.