Phase 1: Core Platform Timeline: TBD Phase 1 delivers the MVP required to operate: a mobile E-commerce app for residents, a web dashboard for pharmacy staff, and the operational backbone connecting orders, payments, inventory, and delivery. 3.1 Resident Mobile App (iOS & Android) A clean, single-codebase mobile application for residents to browse medicines, place orders, track deliveries, and manage refills. Registration: Phone number + OTP . No passwords. Profile stores name, building, and flat number. Medicine catalogue: Searchable by name, generic name, or category. Each item shows dosage, price, availability, and whether a prescription is required. Ordering: Add to cart, upload prescription if needed, confirm delivery address (defaults to saved flat), select payment method. Order confirmation with ETA via app and WhatsApp. Refill reminders: Chronic medications flagged as recurring after first order. Push and/or WhatsApp reminder before expected run-out. One-tap reorder from the reminder. Order tracking: Real-time status: Placed → Preparing → Out for Delivery → Delivered. Notifications at each step. Payments: UPI (GPay, PhonePe, Paytm), debit/credit cards via Razorpay, and cash on delivery. Order history: Full list of past orders with one-tap reorder and downloadable PDF invoices. 3.2 Admin Dashboard (Web) Browser-based dashboard for pharmacy staff and operations managers. Order queue: Live list of incoming and in-progress orders, sorted by time. Audio/visual alert on new orders. Click through to full details including uploaded prescriptions. Order actions: Mark orders as Confirmed, Being Prepared, Out for Delivery, Delivered, or Cancelled. Inventory management: Full stock list with current quantities, reorder levels, and expiry dates. Low-stock alerts. Manual adjustment after supplier deliveries. Expiry flagging. Phase 2: Smart Operations Timeline: TBD Phase 2 introduces the first AI layer, focused purely on operational efficiency. The objective is to reduce stockouts, optimise delivery logistics, and automate the reorder cycle. All models are trained on real data collected during Phase 1. 4.1 Demand Forecasting Predict demand at the SKU level per store using order history, seasonal patterns, and cluster demographics. Approach: Time-series forecasting (Prophet or LightGBM). Simple, interpretable, low compute. T rained on Phase 1 data. Output: Weekly demand forecast per medicine per store, feeding directly into the reorder engine. 4.2 Smart Inventory Auto-reorder: When predicted demand plus safety stock exceeds current inventory, the system generates a reorder recommendation automatically. Expiry risk: AI flags items likely to expire before depletion. Suggests inter-store transfer or promotional pricing. Dynamic thresholds: Reorder levels adjust based on actual demand velocity, not static manual settings. 4.3 Delivery Route Optimisation(if expanded) Batch and route: Multiple orders within a cluster are batched by building/tower and assigned to riders efficiently. SLA enforcement: Orders approaching the 60-minute(requirement chnages*) SLA threshold are automatically prioritised. Real-time re-routing if at risk. 4.4 Operational Analytics The admin dashboard extends with AI-powered insights: Live KPIs: Order-to-door time distribution, SLA compliance rate, stockout rate, repeat order rate, revenue per store. Anomaly detection: Automatic flagging of demand spikes, rider delays, inventory discrepancies. Forecast views: Visual demand and inventory projections for the next 2–4 weeks.