Consumer Attitudes toward Retail Chatbots based on Anthropomorphic Attributes: A Finland–Bangladesh Comparison

Abstract

The aim of this thesis is to evaluate the effects of anthropomorphic attributes towards retail chatbots on consumers' attitudes in two cultural contexts: Finland and Bangladesh. While AI is being integrated more and more into digital interactions to make them more human, the research on the impact of cultural background in users' understanding and reacting to these features is still limited. The study is a comparative qualitative research design with the three factor theory of anthropomorphism. The data collection was based on semi-structured online interviews with six participants (three from Finland and three from Bangladesh) with experience in retail chatbots. The study used two dialogue-based chatbot simulations using ideas from AI. The two chatbots had different interaction styles, one of them was warm and empathetic; the other was direct and professional. Both chatbots were shown to the participants, and they expressed their emotions and ideas about them. They discussed which chatbot they preferred and why. In addition, they described how they would like chatbots to behave in the future. Then the data was analyzed using thematic analysis.

The findings show that the cultural background has a significant impact on the user's reaction to anthropomorphic attributes of the chatbots. The Finnish participants preferred clear, efficient and professional communication. They sometimes considered very emotional language unnatural or inappropriate. For them, trust was largely dependent on the ability of the chatbot to resolve issues in a proper and efficient manner. The people from Bangladesh liked the warm, caring and polite communication. They perceived these features as being caring, respectful, and valued. Trust was built through making a relationship and by making them feel important. Both groups valued responsiveness, although Finns associated it with speed and competence, and the Bangladeshis associated it with attentiveness and care.

The results also indicate that the participants were not only indifferent to one set of chatbot styles. A flexible or hybrid chatbot is a desired option for many of the participants. Empathic and human identity cues like the use of a name were preferred by the participants from Bangladesh, whereas the participants from Finland preferred a chatbot with a human-like personality who can adapt to the situation by adjusting the communication style, depending on the need for efficiency or empathy. Overall, the study explored that there is no single style to be used when designing a chatbot. Rather, chatbots should be flexible and adapt to user needs and context. The results can be used in cross-cultural human-AI interaction studies and provide valuable insights for developing more effective retail chatbots.