Multilingual AI IVR and ERPNext Integration

IVR System + ERPNext Integration for Intercity Bus Operations (South India)

We operate a premium intercity bus service across Tamil Nadu, Andhra Pradesh, Telangana, Karnataka, and Kerala and are building an AI-powered IVR customer care system integrated with our ERPNext instance.

Core requirements:
- Multilingual IVR in Tamil, Telugu, Kannada, Malayalam, and English using Sarvam AI for ASR/NLU/TTS
- Passenger identification via mobile number match or voice PNR entry (our PNRs follow the format ZIEL + digits)
- Self-service handling of vehicle ETA, boarding point info, driver details (masked bridge call), boarding point change requests, and cancellations
- Live vehicle tracking link pushed via WhatsApp/SMS — no verbal GPS readout
- Driver number never exposed; bridge call only within 15 minutes of boarding time
- Automated outbound calls to all passengers on a service when delay or cancellation occurs
- Custom Frappe app on ERPNext with 5 new Doctypes, 11 whitelisted API endpoints, and background scheduler jobs
- WhatsApp-based OCC ticket loop — ops team receives structured alerts, replies "DONE [ticket-id]" to close tickets and trigger passenger confirmation
- Integration with Bitla GDS (read-only), GPS telematics feed (60s sync), and WhatsApp Business API

Stack: ERPNext v14/v15 · Frappe · Sarvam AI · Bolna or Exotel · Bitla GDS API · WhatsApp Business API · Python · REST

Detailed project brief with full IVR flow, Doctype field specs, API endpoint list, SLA rules, and WhatsApp message formats is available after expression of interest.