Abstract
Neuromarketing, a subject that integrates marketing, psychology, and neuroscience to analyse customer behaviour, emerged because of the rapid progress of artificial intelligence (AI) and machine learning (ML) technologies, which have revolutionised marketing. This study investigates how network retail's customer segmentation and conversion prediction may be improved by AI-driven neuromarketing. The main goal was to assess how well predictive models (Random Forest and XGBoost) and clustering algorithms (K-means) identified important customer categories and forecasted conversion likelihood. Using quantitative research methodology, the study examined 50,000 customer records, considering session-specific, behavioural, and demographic characteristics. Features were extracted using Principal Component Analysis (PCA), and the models were assessed using ROC AUC, accuracy, precision, recall, and F1 score metrics. The findings showed that XGBoost performed better in recall (46.67%) and F1 score (0.4882), whereas Random Forest had greater accuracy (60%) and ROC AUC (0.4597). Cluster assignment had little effect on conversion, but PCA-derived features were the most important predictors. These results imply that AI-powered neuromarketing can forecast conversion behaviour and efficiently segment customers, allowing for more individualised marketing tactics. This study advances the industry by showing how cutting-edge AI methods can reveal hidden customer trends, improving network retail marketing effectiveness. Targeted advertising efforts, enhanced consumer interaction, and improved conversion tactics are some of the real-world ramifications.