Remote patient monitoring has become an important part of modern healthcare because it allows doctors to observe patient conditions outside hospitals. However, many existing systems depend on threshold-based alerts or cloud-only processing, which can delay urgent notifications and increase network usage. This paper proposes an AI-enabled remote patient monitoring framework that combines wearable sensors, mobile gateways, edge analytics, and clinical alert prioritization. The system processes heart rate, oxygen saturation, body temperature, and movement patterns near the patient and sends only summarized risk indicators to the healthcare dashboard. A lightweight machine-learning model detects abnormal patterns and assigns risk levels for timely intervention. Simulated evaluation shows 93.1% risk-detection accuracy, 28.4% lower alert delay, and reduced false alarms compared with threshold-based monitoring.