This paper explores how organizations can effectively govern artificial intelligence (AI) in new product development (NPD), focusing on the interplay between formal and informal mechanisms. While AI enhances creativity and efficiency in NPD, its emergent and informal use creates risks related to innovation outcomes. Our study investigates how organizations manage these tensions over time and how governance evolves as AI becomes embedded in NPD.
We conducted a qualitative multiple-case study. Data were collected from 38 semi-structured interviews and supplementary material across 10 companies representing different sizes and products. The data were analyzed through iterative qualitative coding, progressing from first-order concepts to second-order themes and aggregated dimensions (Gioia methodology).
The study develops a dynamic framework of AI governance in NPD comprising four dimensions: AI adoption in NPD, formal mechanisms, informal mechanisms, and dynamic governance for AI in NPD. Findings show that formal controls are typically introduced to mitigate uncertainty but often coexist with informal mechanisms. Effective AI governance emerges from the continuous reconfiguration of enabling and constraining mechanisms rather than from static control.
We conceptualize AI governance in NPD as an evolving configuration of formal and informal mechanisms. By integrating organizational control theory and the Levers of Control framework, it advances a dynamic perspective on AI governance and explains how organizations can manage the tension between innovation autonomy and control in AI-driven NPD contexts.
