As Muscat, Oman undergoes rapid urban expansion with a population reaching approximately 1.74 million in 2024, the demand for efficient emergency services has outpaced traditional manual dispatch capabilities. This paper presents the development and implementation of SmartFort (الحصن الذكي), a fully functional AI-driven emergency management platform developed using R Shiny. The system was designed using the Six Sigma DMAIC methodology to minimize response time and optimize resource allocation across six key agencies: Police, Health, Army, Municipality, Water, and Electricity. SmartFort features a three-tier AI classification engine combining a Large Language Model (Claude Haiku API), a locally-trained Support Vector Machine with Radial Basis Function kernel, and a weighted bilingual keyword engine supporting both Arabic and English. The platform integrates multiple input channels including Google Forms, phone transcription, and SMS for incident reporting. Geospatial optimization is achieved through the Haversine formula for distance calculation and k-means clustering for hotspot detection, enabling predictive deployment of resources. A real-time command-and-control dashboard provides interactive mapping, live performance metrics, incident status tracking, and automated bilingual reporting. The working prototype demonstrates the feasibility of AI-enhanced emergency dispatch in the Gulf Cooperation Council context, with preliminary testing indicating significant potential for reducing triage decision time from 2-4 minutes to under 30 seconds while enabling data-driven resource positioning.

Keywords

Six Sigma, DMAIC, Emergency Management, Artificial Intelligence, Natural Language Processing.