GenAI adoption increases the density of knowledge and collaboration networks: Evidence from a field experiment (Revision 1 )

While generative AI (GenAI) tools can boost individual productivity, their effects on intra-organizational networks for collaboration and knowledge sharing remain largely unknown. The authors theorize that adopting GenAI reshapes employees’ interaction patterns, because it can act as a self-automator (freeing up time for collaboration) and as a knowledge catalyst (increasing the value of individuals as sources of knowledge), thereby rewiring interactions. The authors further propose role-based differences in these effects across specialists (deep domain experts) and generalists (broad integrators). They test these ideas in a randomized field experiment with 316 employees at a European technology services firm, randomly assigning members of 42 teams to either use a GenAI assistant customized with organization-specific knowledge (treatment) or work as usual without it (control). After three months, employees with the GenAI tool became significantly more central in their collaboration and knowledge-sharing networks than the control group, reflecting an expanded set of connections and more frequent knowledge exchange. Specialists experienced greater increases in knowledge centrality than generalists, while generalists gained the most in terms of productivity, measured as number of projects executed during the time window. These findings provide novel causal evidence that GenAI adoption can rewire organizational social structures towards greater connectivity and confirms prior research that it can augment performance differentially.