The rapid emergence of generative AI marks a paradigm shift for knowledge-intensive organizations, but many enterprises still face challenges in translating this potential into tangible value. This thesis explores how generative AI can strengthen knowledge-intensive workflows, drawing on a case study within the product management function of a large enterprise. Using a service design perspective, the research employs qualitative, user-centered methods to identify recurring challenges and evaluate high-impact opportunities for AI intervention.

Findings indicate that while generative AI can streamline tasks such as information synthesis, summarization, and competitor research, lasting value requires more than prompt-based automation. Agentic systems, capable of orchestrating planning, reasoning, and tool use, offer a more robust pathway for integrating AI into real workflows. The development and validation of a agentic research assistant prototype illustrated both the potential and the organizational challenges of implementing agentic AI.

The study concludes that participatory, iterative approaches are critical if generative AI is to successfully augment working practices in knowledge-intensive organizations. By integrating AI into existing workflows and aligning with user needs, organizations can realize its full potential and lay the groundwork for successful AI transformation. This research provides actionable guidance for practitioners navigating the rapidly evolving AI landscape and demonstrates how agentic AI can deliver real value in enterprise contexts.