Initiating and expanding data network effects: A longitudinal case study of generativity in the evolution of an AI platform

Paper in proceeding, 2024

This study explores the emergence and expansion of data network effects (DNEs) in AI platforms. Previous research has focused on direct and indirect network effects. However, the rise of AI platforms necessitates understanding DNEs for platforms' learning and improvement. Through a longitudinal case study of a Conversational AI (CAI) platform's 12-year evolution, the study identifies generative feedback loops as the mechanism for DNEs. These loops are initiated by adding functions that enhance the platform's generative capacity, resulting in more diverse data that improves platform learning. DNEs develop through interactions with different ecosystem actors, including clients and external developers, and rely on various data sources beyond user data to enhance AI platform capabilities. This study contributes to IS literature, specifically digital platform literature, following recent calls to empirically examine DNEs to better understand how AI platforms grow and improve their algorithmic capabilities over time.