Framing Generative Technology for Dynamic Capabilities: A case study of AI platform implementation in large enterprises

Licentiate thesis, 2022

Organizations are increasingly turning to generative technologies known for their inherent dynamic, malleable, and context-agnostic nature to innovate and create a competitive edge. As generative technologies offer virtually unlimited potential applications, organizations are constantly challenged to identify appropriate applications. To date, our knowledge of how organizations implement these technologies to deliver the anticipated outcome is still limited.

Inspired by the recent calls on the nature and management of one such generative technology, Artificial Intelligence (AI), this thesis aims to provide insights on how large organizations make sense of the open-ended nature of AI, and how such framing impacts how they leverage its potential for dynamic capabilities and organizational innovation. As large organizations are typically characterized by established processes, routines, and accumulated collective experiences, this would suggest particularly challenging dynamics when implementing a highly versatile technology.

Following an abductive research approach and a qualitative multiple-case study methodology, this thesis puts forward two empirical papers covering the implementation of an award-winning conversational AI (CAI) platform and its applications, i.e., chatbots and voicebots, across eight large organizations.

Findings indicate that the implementation trajectory differed strongly across those organizations. In all organizations, there were initially inter and intra-organizational incongruent interpretations towards a suitable application of the platform and its use. This illustrates the uncertainty that comes with the open-ended nature of generative technologies, which is in line with prior research. However, contrary to the predominant notions that such incongruencies hinder successful implementation, this thesis illustrates how some organizations actively sought these ‘creative conflicts’ to align diverse perspectives and subsequently uncover new opportunities for dynamic capabilities and organizational innovation. Notably, those organizations shifted from an outcome-oriented to an opportunity-oriented implementation strategy by crafting and employing various cognitive and behavioral processes allowing further exploration of the platform’s generative potential.

Two main practical takeaways can be drawn from this thesis. First, this thesis illustrates that organizations still often evaluate generative technologies using traditional efficiency-orientated key performance indicators (KPIs) that prioritize short-term cost reduction. Such KPIs may be unsuitable for generative technologies that require organizational flexibility to explore the long-term strategic applications related to the ‘horizon of opportunities’ that the generative technology can offer. Second, organizations should be open to rethink their processes, tactics and routines in which they engage in order to realize the full benefits of emerging generative technologies. This is especially relevant for large organizations that have strongly established processes. Findings from this thesis suggest that seeking out and learning from ‘creative conflicts’ is key to adapting those processes, routines, and tactics. This thesis refutes a deterministic view on technology and its implementation. It suggests that organizations must engage in open processes of learning and reframing in order to effectively utilize increasingly malleable technologies.