Exploring frontline ambidexterity in robot restaurants: balancing efficiency, human touch and cultural variations

Purpose This study sought to explore frontline ambidexterity dimensions in robot restaurants, theoretically grounded in the employee–robot–customer service triad and cross-cultural differences between Japan and the USA. Design/methodology/approach The study utilized a mixed-methods qualitative–quantitative observational methodology. It adopted a triangulated approach, comprising three main stages: (1) identification of frontline ambidexterity dimensions, (2) exploration of frontline ambidexterity emergent themes in Japanese and USA service triad and (3) analysis of cross-cultural quantitative patterns of frontline ambidexterity dimensions. Findings Via the multi-stage observational research design, the study identifies four distinct dimensions of frontline ambidexterity: parallel, facilitation, intervention and synergy. These dimensions demonstrate that frontline ambidexterity operates through multiple behavioral mechanisms through which employees balance robot-enabled efficiency and human touch within the service triad. The findings further demonstrate that ambidexterity manifests differently across Japan and USA cultural contexts, with statistically significant differences observed for parallel and synergy behaviors. Practical implications This research offers actionable insights for restaurant managers, HR professionals and service robot designers on how to tailor effective and culturally sensitive robot design, implementation strategies and training programs. Originality/value This study is the first to develop a multidimensional framework of frontline ambidexterity within the employee-robot-customer service triad. The cross-cultural comparison between Japan and the USA further offers a forward-looking perspective on how technological advancement reshapes triadic service dynamics as robotic services evolve from novelty (USA) toward normalization (Japan). Methodologically, this study is among the first to employ a mixed qualitative–quantitative observational approach to examine real-world employee–robot–customer interactions, thereby addressing limitations of prior perception-based studies that could not capture the robot as an active linkage in triadic service encounters.