Abstract
This paper examines how human–machine interaction shapes users’ anthropomorphic perceptions in fintech services. Drawing on an in-depth qualitative case study of an interactive voice response (IVR) system implemented by a microfinance institution in Ghana, we explore how the system’s technological features and design cues afford human-like engagement. Using affordance theory as the analytical lens, we investigate how users interpret, personalise, and act upon the system’s voice persona—known as “Alice.” Based on 154 interviews and extensive observations, the study identifies a typology of perceived social roles through which users relate to the system, a phenomenon we term the “Alice effect.” These include roles such as Caller, Institutional Representative, Monitor, Conversational Partner, and Gendered Persona. The findings reveal how linguistic familiarity, tonal warmth, and cultural resonance trigger both trust and misinterpretation. We further trace how these perceptions influence behavioural outcomes such as savings resumption, branch engagement, and expectation drift. The study contributes to Information Systems scholarship by providing empirical grounding for affective and imagined dimensions of affordance theory in a sociotechnical setting. It also advances understanding of anthropomorphic design in low-trust, low-literacy contexts, offering implications for digital financial inclusion, system transparency, and culturally responsive AI design.