aalto1 untyped-item.component.html
Generative artificial intelligence as a tool for data democratization
Loading...
URL
Journal Title
Journal ISSN
Volume Title
School of Business |
Master's thesis
Electronic archive copy is available via Aalto Thesis Database.
Authors
Date
Department
Major/Subject
Mcode
Degree programme
Language
en
Pages
61 + 15
Series
Abstract
In this thesis a novel way of applying generative artificial intelligence (AI) to enable natural language access to business data has been studied. The thesis aimed to build a generative AI artifact that has been tuned to perform database queries based on user inputs. The research questions and organizational setting has been provided by UPM. To address the given issue an action design research (ADR) approach was adopted.
The research process in this thesis resulted in three main contributions. First, as a result of the ADR process a solution artifact capable of answering 90% of the descriptive and aggregative questions provided to it was built. Second, the research process resulted in a set of three design principles, availability of data, trust towards the artifact and integrability to daily operations. These can be applied to similar projects in the future to guide the development of this kind of artifact. Lastly, the successful development of this artifact enables it to be used as a proof-of-concept (POC) in the case organization, UPM, to decide on further actions and developments.
The research methodology in the study consisted of applying the information technology dominant version of ADR proposed by (Sein et al. 2011). In this process three cycles pre-alpha, alpha and beta, were completed to build the solution artifact. After that the steps of reflection and learning as well as formalization of learnings were conducted.
In addition to studying how the artifact can be built and how well it can be used to access business data, the prerequisites of such artifact were explored. As a result of this several factors like technical and human resources, trust, leadership, data management practices and the data architecture and infrastructure, were identified as important prerequisites.