An End-to-End Customer Intelligence Framework Using Automated MLOps, Behavioral Segmentation, Conversion Prediction, and Prescriptive Analytics for Retail Marketing Optimization

Retail organizations frequently rely on manual data processing and broad marketing strategies that fail to leverage customer behavioral intelligence effectively. This study presents an end-to-end customer intelligence framework designed to automate customer analytics, behavioral segmentation, conversion prediction, and prescriptive decision-making within a unified MLOps architecture. The proposed framework employs a dual-ingestion strategy consisting of an automated watchdog-based monitoring service and an interactive Streamlit interface to support both unattended and user-driven data processing workflows. Customer behavior is analyzed through feature engineering techniques that derive loyalty and lifetime value indicators from transactional records. K-Means clustering is applied to identify distinct customer personas, while an XGBoost classification model estimates subscription and conversion probabilities. The resulting intelligence is translated into actionable business recommendations through a rule-based prescriptive analytics engine. Processed outputs are automatically integrated into a live business intelligence environment through Microsoft Power BI, enabling real-time marketing decision support. The framework demonstrates how machine learning, automation, and business intelligence can be combined to transform raw retail transaction data into targeted and operationally relevant customer insights.