aalto1 untyped-item.component.html
Exploring opportunities and challenges of generative AI in analytics: A data professional’s perspective
Loading...
URL
Journal Title
Journal ISSN
Volume Title
School of Business |
Master's thesis
Authors
Thai, Thao
Date
Department
Major/Subject
Mcode
Language
en
Pages
86
Series
Abstract
Generative AI (GAI) and its rapid advancement has the potential to transform the way people analyse their data. Although there is much promise for GAI, studies have shown that there is a lack of empirical evidence on how GAI can be successfully integrated into common analytics workflows by the users. Therefore, this study explored how GAI was adopted by data professionals using a Technology-Organization-Environment model that was adapted and extended to include a human element to understand the factors that contribute to a sociotechnical transition.
A qualitative methodology was employed for this study, and the data was collected via semi-structured interviews with data professionals in a variety of roles such as data analysts, data engineers, etc. The thematic analysis provided insights on how GAI can offer many advantages, including faster delivery, access to organizational knowledge, and a shift in professional focus from technical menial works to strategic problem-solving. On the other hand, GAI introduces a critical validation burden which creates a unique productivity-reliability trade-off. In other words, the productivity benefits gained from GAI may be offset by the increased cognitive burden of validating the accuracy of the GAI-generated output. In addition to the technical aspects of GAI, the study found that the human aspect is a key factor in determining whether or not one will adopt GAI. Data professionals reported experiencing anxiety and professional identity threat, specifically the fear of losing their skills and becoming obsolete.
The study demonstrated that successful integration would require organizations to go beyond allocating resources to create a symbolic environment that fosters psychological safety. Recommendations for managers include closing the specificity gap in training by designing trainings catered to each specific function, having peers share knowledge and creating designated areas for structured experimentation. This thesis contributed to the body of literature by developing the concept of GAI adoption in analytics by data practitioners as a complex interplay between four distinct but interconnected dimensions: technology, organization, environment, and human.