I’m planning to release a conversational assistant that lives on a mobile device and feels truly helpful, not gimmicky. At its core the app should understand natural language, reply in a tone I can configure, and connect to the usual smartphone capabilities—calendar, notifications, contacts, perhaps even a small knowledge base of company FAQs—so users can actually get things done instead of chatting aimlessly. You are free to recommend the stack. I’m leaning toward a single code-base approach (React Native, Flutter, or similar) but if you believe separate native builds will give a noticeably better user experience, make your case. Likewise, the NLP engine can be OpenAI, Dialogflow, Rasa, or a hybrid; the main requirement is that it can grow from basic Q&A today into richer dialog flows tomorrow. Typical first-release scope • A clean onboarding flow that lets a user set name, preferred language, and notification permissions. • A chat screen with voice-to-text and text input, streaming responses, and quick-reply buttons. • Hooks for three starter skills: scheduling/reminders, FAQ-style info retrieval, and a lightweight support handoff (email or ticket). • Admin console or config file where I can add intents, tweak prompts, and view usage stats. • Source code, build instructions, and a short Loom or Zoom walk-through so I can keep developing after hand-off. Acceptance criteria • App installs on the latest iOS and Android versions with no critical crashes. • Conversational latency stays under two seconds on a decent 4G connection. • Reminders actually trigger local notifications at the correct time. • All user text is stored and transmitted securely (TLS in flight, at-rest encryption in any cloud storage you suggest). If you’ve shipped something similar, include a brief note on the tech choices you made and what you’d do differently this time. I’m ready to start as soon as we agree on milestones and a realistic timeline.