Market readiness for artificial intelligence-enhanced banking services

Abstract

In 2024, AI seems to be everywhere. In November 2022, the first version of ChatGPT was released, and terms that had been perceived as technical and were often misunderstood - like AI, ML, NLP, and neural networks – became ubiquitous. This boosted both development and research, resulting in an avalanche of AI news in all media outlets. While some are excited about everything AI can bring to the table, others are concerned and are trying to raise awareness of the potential risks advanced AI comes with, such as ‘stealing’ human jobs and, in the bleakest predictions, taking over the world. Many are neutral – either due to not knowing enough about AI, or due to not considering it to be a real threat. It is impossible to cover all the ways AI could potentially impact human lives. This thesis focuses on one specific, niche aspect – AI’s impact on business and customer service, and more specifically banking. After an overview of currently available theory and research on AI (from origin and history, to categorization, to use cases in banking), research results are presented and discussed. The research took place in Bosnia and Herzegovina (B&H), an example of a transitional economy. The qualitative portion of the research was done using semi-structured interviews with nine managers from seven banks (representing the majority of the local bank industry). The conversations showed that B&H banks fall into the category of digital laggards. The challenges faced on the way to technological innovativeness were discussed at length (mainly governmental and regulatory constraints). Other aspects of market readiness for AI were discussed, including the perceptions around customer readiness for non-human interactions. The quantitative part of the research builds on top of this. The survey data was gathered from 671 bank customers in Bosnia and Herzegovina. Confirmatory Factor Analysis (CFA) and Covariance-Based Structural Equation Modeling (CB-SEM) were used to analyze the results. Pieces of several popular technology readiness models were incorporated to design the research model used for this thesis. The effects of functional elements, social elements, and technology readiness on attitudes towards AI, and subsequently acceptance of AI, were tested, alongside control variables of education and gender. Testing through 11 hypotheses shows little significance of technological readiness, but important effects of social and functional elements. While gender seems to not have an impact on the results, education matters. The thesis ends with a discussion of results, combining conclusions from theory, previous research, and the qualitative and quantitative portions of this research.