91黑料网

Anand Bhardwaj, a doctoral student at 91黑料网 in the Strategy and Organization area will be presenting his thesis defense entitled:

AI and OR scheduling: Entangling knowledge and artificial intelligence in healthcare

Tuesday, March 24, 2026, at 10:00 a.m.
(The defense will be conducted in hybrid mode)

Student Committee Co-chairs: Professor Samer Faraj鈥�鈥嬧赌嬧赌嬧赌嬧赌嬧赌�

Abstract

Artificial intelligence (AI) technologies in healthcare routinely achieve high levels of accuracy and predictive performance, yet rarely become embedded in everyday care workflows. Understanding this disconnect requires looking beyond technical design to the everyday forms of knowing that make such processes work, and how these forms of knowing are captured in development. Prior research highlights the importance of domain knowledge in AI development but offers limited insight into how multiple situated forms of knowing are negotiated, represented, or excluded as technical systems take shape. This dissertation examines how organizational knowing relates to the development and implementation of a process-specialized AI for operating room (OR) scheduling across two ultra-specialty hospitals developing similar technologies with a common vendor. Drawing on 131 hours of observation, 74 interviews, and in-depth project documentation, I trace how the process of OR scheduling was understood, represented, and enacted during AI development. Each hospital mobilized different local schedule 鈥渆xperts鈥� and thus produced distinct interpretations of what problems could be solved with technology. As a consequence, process-as-represented, formalized in scheduling protocols and project documents, diverged from the cross-occupational process-in-practice enacted by surgeons, clerks, nurses, and administrators. Across both sites, development unfolded through provisionally mobilizing local knowledge, influencing which process knowings-in-practice were included and encoded in the algorithm, and which were excluded. Elements of the AI system grounded in articulated knowings-in-practice materialized and persisted in ongoing work, while those based on excluded forms of knowing dissipated once the projects ended. The study contributes to relational views on technology and organizing by showing that in the context of emergent, negotiated work such as in healthcare, there may be no singular ground truth of the process toward which AI development converges, but rather that AI enacts partial and competing versions of the process. It further contributes to research on knowledge and knowing in organizations by showing how cross-occupational work processes are sustained by the ongoing negotiation of multiple incommensurate forms of knowing, even as technical systems selectively stabilize some while excluding others. Together, these insights explain why many technically excellent AI systems in healthcare fail to take hold: because negotiating knowing is the norm rather than the exception, and most systems are not designed to negotiate.