Madinah Municipality — ACES

Madinah is a holiest-city tier destination with a dense and growing concentration of commercial establishments serving residents and pilgrims. The current manual inspection model covers a fraction of properties, costs significant manpower, and lets violations persist for weeks before an inspector physically reaches them. Documentation is inconsistent; geotagged evidence rarely exists.

Off-the-shelf compliance software doesn't address the scale. The city needed automated, continuous, scalable coverage that produced legal-grade evidence (geotag, timestamp, panoramic photo) without requiring inspectors to be everywhere simultaneously.

The hardware/software boundary is harder than it looks. 360° vehicle-mounted capture has to work in 45°C summer afternoons and dust storms; AI inference has to run cost-effectively on continuous video; the inspector mobile app has to work offline in low-coverage commercial districts.

Drone-compatible 360° camera rigs mounted on inspection-vehicle rooftops capture panoramic imagery as the vehicle drives commercial corridors. Captured frames stream to a cloud AI pipeline that preprocesses, classifies violations (signage, occupancy, permits, hygiene), and tags each one with GPS + timestamp + photographic evidence.

Detected violations land in a Next.js web dashboard with GIS heatmaps, inspector scheduling, and case management. Flagged cases auto-assign to the nearest available inspector through a React Native mobile app with offline-first sync, push notifications, and a clean enforcement-action workflow.

Continuous learning loop: every inspector ruling (confirmed / dismissed / re-classified) feeds back into the model. Accuracy improves with usage, and the city's enforcement patterns shape the AI rather than the other way around.