Description: With AI-based conversational shopping assistance such as Amazon Rufus becoming increasingly more powerful, they carry the potential to make customers’ online experience more personal, engaging and socially rich. However, Rufus appears as a single assistant that aims to help with all questions about all product categories, which contradicts humans’ experiences in the real world. Here, we see specialized product consultants, i.e., different people with skills and expertise. Therefore, we ask the question if different embodiments of the same conversational shopping assistance improve trust, competence, social presence, and purchase intention. To achieve this, the assistant should implicitly communicate its expertise to the customer through affordances.

Goal of Thesis: The goal of this Bachelor/Master thesis (level will determine the scope) aims to study how affordance-based embodiments affect users in their shopping behaviour. This thesis follows a Design Science Research approach and include:

1. Review of the related literature on conversational assistance / agents / avatars / social cues in online shopping, incl. multi-chatbots + identify product categories
2. Solution Design: Conduct a design workshop / generate AI-based designs to produce affordance-based embodied agent designs that are suitable for work tasks.
3. Validation: Evaluate designs in a larger quantitative (online) study.
4. Develop a prototype system with the embodied affordance-based shopping assistant
5. Evaluate how the embodied affordance-based shopping assistant affect customers in their shopping behaviour

Requirements

• Strong interest in embodied AI and online shopping
• Experience with LLMs
• Familiarize yourself with the concepts of Social Presence, Uncanniness, Trust and Competence
• You have read the following papers: [1] [2] [3] [4] [5]