OrderVista

OrderVista is an AI-enhanced logistics dispatch platform that unifies administrative dispatching and driver-side field operations in a single application, purpose-built for last-mile delivery teams and logistics SMEs. ROUTE OPTIMISATION Automatically sequence delivery tasks using a nearest-neighbour algorithm enhanced with Haversine great-circle distance. Mid-shift GPS re-sequencing recalculates the optimal route from your current position after each delivery, and dynamic priority recalculation adjusts task urgency based on time windows and accumulated delays. In testing, route optimisation reduced total delivery distance by 25-31% compared to unoptimised ordering. AI ASSISTANT Get instant AI-powered summaries of your delivery schedule, ask natural-language questions about your tasks, and receive intelligent action suggestions. The agentic AI goes beyond simple chat: it proposes structured workflow actions (mark delivered, go available, sync pending updates) as one-tap execution cards, always with a human confirmation step before anything runs. PROOF OF DELIVERY Capture photographic evidence and scan barcodes or QR codes directly from the app when completing a delivery. Photos, scans, and driver notes are securely uploaded and linked to each task for administrative review. FLEET MANAGEMENT (ADMIN PORTAL) Dispatchers access a full web-based admin portal for task creation, CSV bulk import with automatic geocoding, driver assignment with performance-aware recommendations, fleet-wide distance minimisation across multiple drivers, proof-of-delivery review, customer tracking code generation, and notification dispatch. REAL-TIME COMMUNICATION Built-in live chat connects drivers and dispatchers instantly for operational coordination, reducing reliance on external messaging apps. DRIVER RECOMMENDATION A composite scoring engine evaluates driver availability, GPS proximity, current workload, historical completion rate, and average delivery time to recommend optimal task assignments. OFFLINE-FIRST ARCHITECTURE Continue working even with poor or no connectivity. Completed deliveries and status updates queue locally and sync automatically when your connection returns, using progressive exponential backoff to avoid flooding the server. MULTI-TENANT SECURITY Enterprise-grade data isolation enforced at the database level through PostgreSQL Row-Level Security. Each logistics company's data is completely separated with no application-side middleware required. CUSTOMER TRACKING Auto-generated tracking codes let customers check delivery status through a public lookup endpoint, keeping them informed without needing an account. TECHNOLOGY Built with Flutter for cross-platform deployment, backed by Supabase (PostgreSQL, Auth, Realtime, Storage), with Mapbox for geocoding and maps, and DeepSeek R1 via OpenRouter for AI capabilities. Developed as a Final Year Project (CS4514) at the City University of Hong Kong.