From grand visions to situated imaginaries: AI adoption under constraint in retail SMEs

This article develops the concept of situated imaginaries of constrained automation to explain how the consequences of AI adoption are imagined by actors from small and medium-sized enterprises (SMEs) operating under sociomaterial constraints. Drawing on the framework of sociotechnical imaginaries and insights from critical AI studies, we shift attention from macro-level policy visions to everyday organisational contexts in which AI adoption is shaped by limited time, labour, capital, expertise and infrastructural autonomy. The analysis is based on qualitative research conducted in the European retail sector, where AI uptake remains uneven and resource-dependent. Using reflexive thematic analysis of co-creation workshops with retail SME representatives, we identify a recurring pattern of future-oriented, interrelated narratives structured by interlocking organisational, resource and infrastructural constraints: AI as organisational compensation, AI as incremental displacement and AI skills as a stabilising solution that reframes structural pressures as questions of adaptability. By conceptualising imaginaries as situated and constraint-bound, the article offers an empirically grounded contribution to research on sociotechnical imaginaries and to organisational studies of digital transformation.