Organizations from all industries have recently begun to develop and operate machine learning (ML) systems. While ML promises to improve an organization's effectiveness and efficiency, developing and operating ML systems remains challenging as these systems differ significantly from traditional software and require novel work practices that run counter to existing business processes. These conflicting demands create tension in the organization as resources to develop and operate ML systems are limited. Organizations thus seek to leverage scarce resources by employing a range of organizational structures and tailored tactics. To explore the interplay of organizational structures, tensions, and tactics, we conducted an explorative expert interview study informed by computational grounded theory methodology. We took an ambidextrous perspective to identify four central tensions and associated tactics employed within given organizational structures. Further, we found that organizations are moving from centralized and decentralized structures to hybrid ones to enable effective ML development and operation.


Paper Number 1230; Track Governance; Complete Paper



When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.