I’m building a round-the-clock voice assistant that can speak with passengers just as a human airline representative would. The agent must: • Hold natural, back-and-forth conversations in clear, idiomatic English, recognising a wide range of traveller queries without sticking to rigid scripts. • Complete core airline tasks end-to-end: make new flight bookings, add ancillaries, manage check-in, push real-time status updates, and guide customers through rebooking or cancellations when plans change unexpectedly. • Personalise every response by drawing on passenger and flight data pulled securely from our existing reservation, loyalty and operations APIs. • Stay available 24/7 across whichever channels you recommend—mobile, web, or smart-speaker integration—while remaining scalable during peak travel periods. Deliverables I need to see before sign-off: 1. A working voice interface (Twilio Voice, Dialogflow CX, Amazon Lex, Rasa, or similar) trained and tuned for airline vocabulary. 2. Secure middleware that connects the agent to our booking, check-in and flight-status endpoints with proper authentication and logging. 3. A testing suite demonstrating successful dialogue flows for booking, check-in, disruption rebooking and cancellation scenarios. 4. Deployment scripts and concise operational documentation so my in-house team can maintain, retrain and expand the model. If you’ve previously shipped voice bots for travel or another highly regulated sector, I’d love to review a short demo or case study. Please outline your preferred tech stack, timeline, and any additional NLP enhancements—such as accent handling or sentiment recognition—you believe would strengthen the experience.