Empowering Retail Visitors Through Edge AI: A User-Centered Case Study of Cooperative Totem Systems

Retail is one of the most significant and competitive economic sectors, and interactive systems such as informative totems play an increasingly important role in enhancing the visitor experience within shopping mall environments by offering value-added services. Over the years, these systems have evolved toward AI-driven context awareness to personalize content based on the people interacting with them. To maintain soft real-time responsiveness, however, existing systems typically rely on cloud-based processing of AI workloads. This raises significant privacy concerns, as sensitive visitor data must be transmitted to external systems or third-party services. Although prior literature demonstrates notable progress in integrating AI-driven context awareness into interactive systems for shopping mall environments, current solutions do not enable fully edge-confined execution of AI workloads and therefore cannot guarantee that sensitive data remain local. To address this gap, this paper presents an industrial case study aimed at developing a Cooperative Intelligent Totem System, in which all AI tasks are executed locally within an edge infrastructure composed of a cooperating totem and roof nodes, without relying on any external systems or third-party services. Experimental results show that the system achieves accurate AI-driven perception, consistently satisfies the one-second responsiveness requirement, and scales up to 18 simultaneous users when supported by multiple cooperating roof nodes, all while keeping sensitive data strictly confined to the edge.