Scaling AI Ventures: How to Navigate Tensions between Automation and Augmentation

Abstract

AI ventures promise to automate and augment ever more human tasks. This provides rich opportunities for growth. Yet, digital and human resources that involve AI are oftentimes task-specific and hard to scale. Furthermore, clients remain skeptical to be fully automated by external services. Thus, it remains unclear how AI ventures achieve growth. We adopt a grounded theory approach on an interview study with founders, product managers and investors to inquire how resources afford or constrain scaling in AI ventures. For this, we blend the notion of (non-)scale free resources with the layered architecture of digital technologies. Our study suggests that AI ventures scale by organizing digital and human resources for replicability in that they keep AI-specific resources distant from clients while simultaneously externalizing human-intensive tasks to their clients. As we inquire the roles of human and digital resources, our study suggests that ventures seek to quickly find an optimal degree on the continuum between augmentation and automation when bundling resources.