From divergence to convergence: GenAI's role in collaborative design thinking

Generative AI (GenAI) offers significant potential to boost innovation performance by accelerating speed and productivity. However, its integration into collaborative, empathy-driven processes like design thinking remains a complex challenge. Drawing on qualitative data from three innovation sprints involving 71 interdisciplinary teams and 18 corporate partners, we investigate how GenAI impacts the core design phases of problem understanding, ideation, and prototyping. Our findings demonstrate that while GenAI effectively expands the problem space and generates diverse solutions, its integration varies significantly across teams. We identify a critical tension: although GenAI can enhance early-stage divergence, its output often lacks the contextual and empathic depth required for user-centered innovation. To address this, we propose a 2×2 framework defined by two dimensions – Level of Integration (High vs. Low) and Purpose of Use (Supportive vs. Directive) – identifying four distinct engagement patterns. This typology illustrates the trade-offs between productivity and team-based empathy and coordination, providing a nuanced perspective on how AI reshapes different stages of the design process. Finally, we offer practical guidance on team orchestration and intervention strategies to ensure GenAI augments, rather than undermines, collaborative creativity.