Smart Delivery Intelligence Platform
Swiftlane is a proactive last-mile delivery failure prevention engine. It predicts delivery failures before they happen using real-time signals — weather, traffic, and customer history — and autonomously resolves them via WhatsApp with zero dispatcher intervention.
Last-mile delivery accounts for 53% of total logistics costs. Every failed delivery attempt costs ₹80–200 in re-delivery fees. Current systems are reactive — they log failures after they happen. Swiftlane fixes this.
- Every delivery gets a Failure Risk Score (0–100%) calculated from live weather, traffic, and customer data
- If the score crosses 75%, an automated WhatsApp message is sent to the customer
- The customer's reply is parsed by AI — driver instructions update instantly
- Dispatchers monitor all deliveries on a live risk heatmap in real time
| Layer | Technology |
|---|---|
| Frontend | Next.js + Tailwind CSS + Mapbox GL |
| Backend | Node.js + Express |
| ML Service | Python + FastAPI + scikit-learn |
| AI / NLP | OpenAI API + LangChain |
| Messaging | Twilio (WhatsApp Business API) |
| Database | PostgreSQL + Redis |
| Infrastructure | Docker + GitHub Actions |
├── frontend/
├── backend/
└── ml/
- Risk Brain — ML model scoring every delivery for failure probability
- Autonomous Triage — AI-powered WhatsApp conversation flow triggered automatically
- Live Dashboard — Mapbox heatmap with real-time delivery status and risk indicators
- Proof of Presence — GPS geofencing + timestamped photo before marking delivered
Setup instructions per service coming soon. Each service runs independently — see individual folders for details.