Understanding Knowledge Work and the Performance Potential of its Computerization. Case IBMs Watson

Abstract

This study focuses on theorizing knowledge work and studying the performance potential of computerizing contemporary knowledge work tasks. In the research, the job content of five knowledge workers is described, classification for knowledge work task is formulated, and a framework for knowledge capabilities required in performing knowledge work tasks is constructed. Furthermore, the technological properties of IBM's new questions-answering computer, Watson, are described in general and its knowledge capabilities and performance potential in knowledge work tasks are analysed.

The literature review of this research concerns the nature of information and knowledge, knowledge work, and knowledge work performance and productivity. Moreover, two models on cognition are presented that help understanding the mind. Despite the fact that knowledge work has attracted numerous scholarly minds, no clear and concise definition exists. The research was conducted using the methods of qualitative case study and grounded theory. Qualitative case study has been used to describe and explain the multifaceted phenomenon. Grounded theory method has been applied in constructing the framework in the study. Empirical evidence is based on the interviews of five knowledge workers and a technology expert as well as on relevant published data on Watson. Inductive content analysis has been used to study the interview material by categorizing the jobs of the knowledge workers into roles and tasks. The Spaun model and COGNET framework presented in the literature review have worked as intellectual guides in constructing the knowledge agent’s knowledge capabilities framework applied in the analysis. Descriptions on Watson’s technology and capabilities are based on published material and its knowledge capabilities have been analysed by using the knowledge capability framework.

Among the key results of this study are the formulation of knowledge work task typology, the construction of framework for knowledge capabilities, and the analysis of Watson's performance potential in various knowledge work tasks types. The findings of Watson's performance potential analysis suggests that it has the greatest performance potential in the task types of answering questions, using analyzing tools to create information and insights, information disseminating, requesting information and delegating. Its lowest performance potential is in the task types of directing discussion, generating ideas and alternative solutions, persuading and negotiating, and discussing and deciding together, according to the results.