I’m setting up an AI-driven voice agent dedicated to customer support and want it to feel as natural as speaking with a live representative. The core focus is twofold: • Answer routine FAQs so callers get instant, accurate information without waiting in a queue. • Process straightforward orders from start to confirmation, then hand off complex cases to a human when needed. The agent must converse fluently in English, Urdu and Hindi, switching languages automatically once it detects the caller’s preference. Crystal-clear speech recognition, natural-sounding TTS, and intent handling via GPT-style NLP are all expected under the hood. I’ll provide sample call flows and knowledge-base content; you structure the dialog, train the model, and deploy it to a phone line or SIP endpoint I can test from day one. Deliverables 1. A fully functional voice bot accessible by phone/SIP. 2. Source code, dialogue design files and setup documentation. 3. Quick-start guide so my in-house team can update FAQs and product lists without touching code. Acceptance criteria: callers can complete at least 90 % of trivia FAQ questions and simple product orders end-to-end in any of the three languages, with a fallback to live staff on errors. If you’ve implemented Twilio Voice, Dialogflow CX, Amazon Connect, or similar stacks before, that experience will be a plus—please highlight it.