The Impact of Product Review on Digital Twins’ Choice: Convergence, Divergence, and Predictive Alignment with Human Consumers Judgments
(A2026-133898)

AUTHORS

Kate Jeonghee Byun, NEOMA Business School; MICHAEL HAENLEIN, ESCP Business School/University of Liverpool Management School; Daegon Cho, KAIST

ABSTRACT

Firms are increasingly adopting large language model based digital twins that simulate individual consumers by internalizing their preferences and decision cues. Yet it remains unclear whether these twins truly reflect the choices of the human consumers they are designed to represent, and under what conditions such alignment improves or deteriorates. This research develops a conceptual and methodological framework to examine the role of digital twins as consumer proxies. We further investigate whether digital twins, constructed using self-reported survey responses and real purchase logs, exhibit choice patterns comparable to those of humans in terms of predictive accuracy. Additionally, our framework tests how key review characteristics, spanning psychological, contextual, and informational dimensions, shape the conditions under which human and digital twin choices converge or diverge. This work provides foundational guidance for understanding when digital twins behave like their human counterparts and how alignment is influenced by factors such as review structure, emotionality, and reviewer identity.