AI-powered voice assistants have evolved into sophisticated agents capable of social and emotional intelligence. This paper conceptualizes a new class of AI agents, empathic voice assistants (EVAs), which can detect, interpret, and respond to users’ emotional states in real time. In voice commerce, EVAs represent a novel intervention that could reduce the gap between consumers’ pro-environmental intentions and purchasing behaviors. While consumers express concern for sustainability, they often default to habitual or impulsive choices. By adapting to individual preferences and emotions, EVAs can timely and gently guide consumers toward more sustainable alternatives. We propose that EVAs, when functioning in various roles—advisor, educator, coach and companion—can serve not only as shopping assistants but also as sustainable behavioral guides. Building on recent advancements in affective computing and AI empathy, we present eight conceptual propositions that establish a theoretical foundation for understanding how empathic AI can shape sustainable shopping decisions. We conclude by discussing key implications for managers and policymakers, outlining attention points for the responsible and effective deployment of EVAs in retail environments. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.