Large language model tools as catalysts for collective cognition in collaborative new-product development: a quasi-experimental study

Modern organisations frequently face complex cognitive challenges in Collaborative New-Product Development (Co-NPD), particularly when integrating dispersed knowledge and coordinating work across different development phases. To investigate how Large Language Models (LLMs) influence collective cognition and collaborative processes, this study introduces the 2I2A model, which defines four collaboration spaces and eight associated communication dimensions. A mixed-methods design was adopted, combining quantitative analysis of collaborative behaviours with semi-structured interviews analysed using grounded theory. The findings indicate that LLMs mainly support the early stages of Co-NPD by expanding collective cognitive boundaries, improving knowledge integration and facilitating idea generation. However, their contribution to deeper analytical reasoning, negotiation and solution integration is more limited. The grounded theory analysis additionally highlights potential drawbacks, including reduced collaborative naturalness and a tendency toward over-reliance on LLM-generated suggestions. Overall, the study suggests that the 2I2A model offers a useful framework for examining how collective cognition develops in Co-NPD and clarifies both the potential and the boundaries of LLM assistance in collaborative innovation.