AI-Powered Wellbeing Platform

About the Client

Phi AI is a mental wellbeing and emotional support platform created by a professional coach to support busy working parents navigating stress, burnout, and emotional overload. The client's vision was to offer daily, judgment-free, coach-like support that feels accessible, private, and emotionally intelligent—without replacing therapy or clinical mental health services.

The target audience includes high-achieving professionals and parents who often carry emotional weight silently, struggle with people-pleasing behaviors, and seek clarity, relief, and intentional change in their daily lives.

Visit Phi AI

Project Overview

Phi AI is a cross-platform AI-powered wellbeing system consisting of a mobile application (iOS & Android) for end users, a secure backend responsible for AI orchestration and privacy enforcement, and a web-based Admin Dashboard for monitoring, moderation, and analytics.

The system integrates Large Language Models (LLMs) to deliver personalized, empathetic, coach-style conversations. Our responsibility covered end-to-end product development, including architecture design, AI integration, privacy-first authentication, mobile UX, backend services, and admin tooling.

AI-Powered Conversational Coaching

• Natural, supportive chat experience inspired by professional coaching sessions
• Context-aware responses that evolve over time
• Emotion-sensitive tone aligned with the user's mental state
• Multilingual support (English & French)
• Voice-to-text input using AI speech recognition

Privacy-First Anonymous Experience

• Anonymous, token-based authentication (no email, no username, no personal data)
• Users identified only by randomly generated backend tokens
• Optional biometric protection (Face ID / Touch ID)
• Automatic data retention limits to minimize long-term storage

Emotional Check-Ins & Pattern Recognition

• Free-text or voice conversation input
• One-tap emotional check-ins (e.g., sad, anxious, happy)
• Emotional trends detection by time of day or week
• Pattern analysis for frequently discussed topics

Admin Dashboard

• View anonymized users and conversation history
• Monitor sentiment distribution (positive / neutral / negative)
• Track user engagement and usage patterns
• Manage and refine the LLM system prompt