Artificial Intelligence (AI) has quietly rewritten the rules of online retail, and nowhere is this more visible than in the rise of intelligent shopping assistants that chat, recommend, and reassure customers in real time. A growing body of work now treats “friendliness” warmth, empathy, responsiveness, and human-like conversational style as a design variable in its own right, distinct from raw functional performance (Cheng et al., 2022; Singh et al., 2024). This conceptual paper asks how friendly AI shopping assistants shape customer trust and, through that trust, sudden or unplanned (impulse) buying in online retail. Drawing on the Stimulus–Organism–Response (S-O-R) framework (Mehrabian & Russell, 1974), the Technology Acceptance Model (Davis, 1989; Venkatesh et al., 2003), Social Response Theory (Reeves & Nass, 1996), and Trust Theory (Mayer et al., 1995), the paper proposes that AI friendliness functions as a stimulus that heightens social presence, emotional connection, and perceived credibility. These psychological responses, in turn, strengthen customer trust and raise the likelihood of impulse purchasing (Rook, 1987; Li et al., 2022). The paper argues that existing literature has treated AI friendliness, trust formation, and impulse buying largely as separate streams (Blut et al., 2021; Hu et al., 2019), and it proposes an integrative conceptual model and four testable hypotheses to close that gap. The paper contributes to AI-retailing scholarship by positioning trust as the central mechanism linking friendly AI interaction to consumer buying behaviour, and it closes with managerial guidance on balancing personalization, transparency, and ethical persuasion, along with directions for future empirical validation using structural equation modelling.