An AI-driven chatbot must be built to handle customer support around the clock on both our website and companion mobile app. It should greet visitors, answer common questions, guide them through troubleshooting, and seamlessly transfer more complex cases to a live agent when necessary—all in English. Scope • Architect, train, and deploy the model so it understands our product range, policies, and tone of voice. • Embed the chat interface on our website (JavaScript snippet or widget) and integrate the same bot into the mobile app via SDK or REST API. • Provide an easy-to-update knowledge base so my team can add new FAQs without touching code. • Include basic analytics: conversation volume, resolution rates, and missed intents. Acceptance criteria • 90 %+ intent recognition accuracy in English test set supplied. • Under 2 seconds response latency on both platforms. • Handoff to human agent triggers correctly in >95 % of flagged cases. • Installation instructions and source (flows, training data, integration code) delivered in a Git repo. Open to proven frameworks such as Dialogflow, Rasa, Microsoft Bot Framework, or a custom LLM approach if you can meet the above metrics. Please outline your preferred stack, timeline, and any similar deployments you have completed.