Exploring the collaboration between humans and AI-grounded epistemic objects to facilitate knowledge integration of “who knows what” under corporate crises

Purpose This study aims to explore how human actors collaborate with artificial intelligence (AI)-grounded epistemic objects (EOs) to facilitate knowledge integration of “who knows what” during corporate crises. Design/methodology/approach This study used a 15-month focused ethnography at an internationally renowned Chinese organization. Findings Human actors collaborate with AI-grounded EOs to facilitate knowledge integration of “who knows what” during corporate crises through connecting, eliciting, committing, converging, combining and applying across three iterative stages of specialization, credibility and coordination. Originality/value This study offers three novel advances: (1) through a 15-month focused ethnography, the authors reveal how human intelligence (HI) and AI synergize to enable knowledge integration via a three-stage mechanism involving six micro-processes, uncovering the “black box” of AI–HI collaboration. (2) The authors uniquely apply EO theory to AI-generated outputs in knowledge management–corporate crisis management contexts, demonstrating the agentic role of AI-grounded EOs in strategic knowledge synthesis, beyond its typical framing as passive “tools”. (3) The model resolves tensions in balancing algorithmic rationality with human moral judgment during high-stakes crises.