Our web portal and its companion mobile app need an AI-driven customer-support chatbot that does more than recite scripted answers. I want a solution that listens, learns, and adapts—tracking user intent over time, refining replies, and surfacing helpful resources automatically. Key expectations • The bot must live natively on both the website and the iOS/Android builds, sharing a single knowledge base so conversations can continue seamlessly across devices. • Natural-language understanding, sentiment detection, and continual learning are essential; think GPT-powered dialogue management or a comparable Rasa / Dialogflow CX stack with custom machine-learning layers. • It should integrate with our existing support ticket API to open, update, and close cases, while logging every interaction for later analysis. • A simple training dashboard is needed so my team can add new FAQs, review conversation logs, and fine-tune intents without touching code. Deliverables 1. Production-ready chatbot deployed to staging and then live environments 2. Source code, model files, and infrastructure scripts (Docker/Kubernetes or similar) 3. Admin/training dashboard with role-based access 4. Setup documentation and a brief hand-off workshop The project will be accepted once the bot consistently resolves at least 70 % of common support queries unaided during a monitored test week and demonstrates measurable improvement from user feedback.