AI-Driven CRM Architecture for Managing Large-Scale Fragrance Sample Requests and Understanding Customer Preferences on Social Media

Social media platforms have become critical infrastructures for customer relationship management (CRM), requiring scalable and intelligent solutions to handle high-volume interactions. In the luxury fragrance sector, digital promotion poses a unique challenge because olfactory attributes cannot be experienced online. As a result, physical fragrance samples remain essential, generating large volumes of sample requests or inquiries across social media. However, many requests remain unmanaged due to limitations in manual CRM (i.e., human-driven processes), revealing a design gap that may negatively affect perceived responsiveness and service quality. This study uses qualitative content analysis with NVivo 12 to examine large-scale sample request interactions on the Facebook pages of four luxury fragrance brands. Data was collected via NCapture and analyzed to identify recurring patterns, linguistic structures, and customer expressions related to sample requests. Findings confirm frequent repetitive requests, highlighting inefficiencies in traditional CRM systems under high demand. This research proposes an AI-driven CRM Sample Request Management Architecture (CRM–SRMA) that systematically captures and processes customer sample requests, collects the necessary mailing information, and seamlessly transfers validated data to the final dispatching stage. The proposed system also models individual fragrance preferences by analyzing customers’ interactions with samples, particularly in terms of top, middle, and base notes. By leveraging this information, the architecture enables the targeted promotion of new fragrance releases that closely align with customers’ demonstrated olfactory preferences. The insights of this research provide a scalable, intelligent mechanism that enables luxury social media managers and CRM systems to manage high-volume interactions while maintaining service quality. By automating sample request processing, the mechanism improves responsiveness and reduces operational burden. It also supports long-term relationship building through preference tracking and updating customers with any new relevant-fragrance releases. Although focused on fragrances, the mechanism is adaptable to other luxury cosmetic categories, thereby ideally enhancing overall social media-based customer service.