From Smart Assistants to Smart Spaces: AI in Fashion Retail and the Emerging Need for In-Store Behavioral Intelligence

In the context of the digital transformation of the retail sector, physical fashion stores are increasingly seeking intelligent systems that can provide real-time insights into shopper behaviour. This paper presents an AI-powered monitoring system that has been deployed in a large fashion retail store. It integrates a dense network of Xovis top-view sensors with a real-time analytics. The system uses a temporal proximity-based clustering algorithm to perform fine-grained analysis of customer flow, gender composition, and group segmentation. Over the course of one year, the system recorded and analysed more than 479,000 shopper interactions, revealing temporal patterns, spatial preferences and behavioural trends. The results demonstrate a consistent predominance of female shoppers, a high frequency of individual shopping behaviour and distinct peak hours that align with daily routines and promotional campaigns. The analytics are designed to support operational decision-making and enables managers to optimise store layout, staffing and marketing strategies. The proposed system is a scalable, privacy-conscious solution for behavioural intelligence in physical retail. It offers a foundation for predictive modelling, adaptive merchandising and data-driven retail design. Future developments will explore extended behavioural profiling, automated alerts and integration with business KPIs to enhance strategic decision-making.