I’m building a customer-facing AI Voice Agent that can speak to our skincare shoppers any hour of the day and give genuinely personalized guidance. Here’s what I need built and delivered: • Cross-channel integration – The agent must live inside both our website and our iOS/Android app, sharing one knowledge base so a user can move seamlessly between channels. • Hybrid voice experience – In conversation it should blend high-quality text-to-speech with select recorded human snippets for greetings and key brand moments, switching automatically to keep the tone warm yet scalable. • Smart personalization – Recommendations must factor in each visitor’s stated skin type, current concerns, and the routine they already follow. The agent should store these data points securely and reference them in future chats. • Dynamic expertise – I want a simple admin panel or API hook that lets my team feed in new product launches, ingredient studies, and evolving trends so advice stays current without redeploying the model. • Core skills – Product matching, step-by-step routine building, usage clarifications, and troubleshooting common issues—all delivered conversationally, with fallback to help articles when needed. • Always-on reliability – Latency under two seconds for voice responses and uptime suitable for 24/7 support. Please build a functional prototype that includes conversation flows, voice synthesis setup, and integration stubs for both web (React) and mobile (Flutter). I’ll supply product data and brand recordings; you bring the NLP/voice tech, dialog design, and secure user data handling. A quick video demo plus deployable code in GitHub will be considered a complete first milestone, followed by two revision cycles to polish the experience.