Hiyori - You Health Care Agent

Health Care Chat Box Layout
Admin Function 1
Admin Function 2
Admin Function 3

🌱 Building a Serverless AI Health Agent to Support Hong Kong’s Youth Mental Well-being Hiyori – Your Health Care Agent

Introduction Youth mental health in Hong Kong is facing an unprecedented crisis. In 2023, student suicides reached a decade-high of 32 cases—nearly triple the number from ten years ago. Surveys indicate that over 50% of secondary students experience depressive symptoms, and more than 80% report anxiety about their future. Despite this, nearly half of caregivers hesitate to seek professional help due to stigma or lack of awareness. To address this urgent gap, our team developed Hiyori, a serverless AI-powered health agent designed to support emotional well-being among children and teenagers, while empowering caregivers with timely insights. Built on AWS cloud infrastructure, Hiyori offers a scalable, stigma-free solution tailored to Hong Kong’s cultural and linguistic context.

Problem Statement • High prevalence of ADHD, anxiety, depression, and disruptive disorders among youth. • Suicidal ideation reported by nearly 7% of students in 2023. • Stigma and lack of awareness hinder help-seeking behavior. • School counselors are overwhelmed, with mental health dominating case loads. Hong Kong urgently needs a tech-enabled, emotionally intelligent solution that meets youth where they are—digitally and emotionally.

Solution Overview Hiyori is a multilingual, serverless AI companion designed to detect emotional distress, offer coping strategies, and facilitate communication between children and caregivers. Key Features • Children’s Chatbox: A conversational interface in Cantonese and Traditional Chinese that allows children to express their emotions. The AI detects signs of distress and recommends coping tools such as breathing exercises or mindfulness videos. • Parental Guidance: Caregivers receive timely notifications and suggestions to better understand and support their child’s emotional needs. • Safety First: Hiyori does not provide medical diagnoses. Instead, it focuses on empathetic dialogue, early detection, and escalation workflows when necessary

User Story: Meet Alex To show how this works in practice, let’s look at Xiaoming, a secondary school student in Hong Kong. “Lately I feel very anxious because of too much homework. I’m scared I’ll fail the exams. I often feel that I’m not good enough, and I can’t do anything right. Sometimes I really don’t want to go on anymore.” When Alex shares these feelings through the AI Health Agent, the system detects signs of anxiety and low mood. It gently guides him to try a breathing exercise video, while at the same time sending a discreet notification to his parent encouraging them to provide extra care and support. This example demonstrates how the AI Health Agent can help students express emotions safely, give them practical coping strategies, and ensure parents are alerted early if there are warning signs.

User Story: Meet Jamie To show how this works in practice, let’s look at Jamie, a university student living in Kowloon. “Recently I’ve been feeling really frustrated about my weight. I’ve tried different diets but nothing seems to work. I feel tired all the time, and I’m losing confidence in myself. I just want to feel healthy again, but I don’t know where to start.” When Jamie shares these concerns with the AI Health Agent, the system identifies signs of weight-related stress and low motivation. It gently recommends a personalized meal plan and a beginner-friendly workout routine tailored to Jamie’s lifestyle. At the same time, it sends a supportive message to Jamie’s close friend, encouraging them to offer companionship and positive reinforcement. This example shows how the AI Health Agent can help users tackle physical health challenges with empathy, provide actionable advice on diet and exercise, and activate a caring support network to boost long-term well-being.

Healthcare AI System Administration & Testing Platform Overview The highly integrated Healthcare AI - Admin Dashboard empowers system administrators (such as Dr. Sarah Li) to oversee the full operation of the AI-powered healthcare support system. Through a suite of tools, administrators can analyse user data, monitor conversation records, and identify early signs of potential mental health risks.

User Profiles & Conversation Monitoring Via the Demo User Profiles panel, administrators can access simulated user characteristics and interests—for example, teen user Alice Chen (interested in anime, gaming, and music; energetic, curious, and social) and elder user Margaret Wong (interested in gardening, cooking, and history; wise, patient, and nurturing). These profiles help simulate realistic interaction scenarios and allow the AI to tailor its responses based on personalized traits. Within the Demo Users panel, the system displays specific chat test results. These conversation logs not only reveal the AI’s response logic but also enable administrators to assess whether users are exhibiting signs of anxiety, depression, or other emotional distress.

Chat Testing & Early Risk Detection Through the Chat Test functions—such as “Test Teen Chat,” “Test Elder Chat,” and “Test AI Responses”—administrators can simulate various dialogue scenarios to observe whether the AI accurately identifies emotional cues and delivers appropriate responses. This helps determine whether a user should be referred to mental health professionals. These tests serve not only as technical validations but also as critical tools for early intervention. For instance, if the AI detects that a teen user expresses significant anxiety or loneliness, the system can automatically flag the user as “requiring further attention” and notify the relevant healthcare team for follow-up support.

Impact This project directly addresses three pain points:

Early Detection: The AI agent lowers the threshold for kids to express distress.
Family Support: Parents get actionable insights without stigma.
Scalability: Schools and NGOs can adopt the solution cheaply, since serverless services scale automatically with usage. Our long-term vision is to collaborate with schools and NGOs to reduce barriers to help-seeking and provide every child in Hong Kong with a safe, tech-enabled companion for mental well-being.

Conclusion Hong Kong’s youth are facing serious mental health challenges—but technology offers a beacon of hope. By combining AWS serverless infrastructure with AI safeguards tailored to the unique contexts of both Hong Kong and Mainland China, we’ve built a secure, scalable, and cost-effective solution that empowers children, supports parents, and strengthens community cohesion. This project is more than just code—it’s about giving children a voice and helping families find a path forward. In a short period of time, we designed and deployed a functional prototype that integrates AI and cloud services to address a pressing social issue. Hiyori is not merely a tool; it’s a scalable and compassionate mental health solution. As it continues to evolve, it holds immense potential for commercialization and social impact—both in Hong Kong and around the world.

TRY IT OUT !

http://healthcare-ai-frontend-1759164955.s3-website-ap-southeast-2.amazonaws.com/