Kidddulting
An app product designed for the psychological healing of the "inner child" through multimodal AI Technology.
#Industry Project #Experimental #Team Project #Mobile #Multimodal AI
Roles
In the team, I am the Product Manager leading the process of user research, ideation, technical support and UI/UX design. Other members:
Hairu Zhai: UI/UX Designer
Ziyi Wang: User Researcher
Tian He: Storytelling
Problems
In the “Kidulting“ social trend, the emotions of the “inner child” are latent and not easily found for adults today. And the outside supports from family, friends and doctors are limited to this psychological phenomenon.
Techniques
User Interview, KANO Model Analysis, Product Requirement Document (PRD), Wireframing, Prototyping, Usability Test, Visual Programming
Objective
A therapeutic app designed to help users find their “inner children" through talks with AI characters, guiding people to discover hidden emotions and release the internal pressure.
01 Context
“Kidulting” describes a social trend in which adults participate in activities originally designed for children.
This is a neologism that combines the words “kid” and “adult”. “When I was a kid, I wanted to be an adult; when I grow up, I want to be a kid” shows the psychological phenomenon among adults today.
It is a way helping people to short escape from the hard reality and acquire support and encouragement to face life challenges.
02 Secondary Research
The emotions of the “inner child” are latent and not easily found while the external supports from family, friends and doctors are limited.
We drew two conclusions from the publication by an psychological expert, Qijia Shi:
A research paper titled: “Predictors of help-seeking behaviour in people with mental health problems: a 3-year prospective community study“ by Doll, C.M., Michel, C., etc. shows that in the sample group:
63.1% would first seek help from family or friends when facing psychological stress;
28% would first seek help from professional therapists.
Among them are still concers:
At this stage, we understand that peole need the therapy which is always-available, privacy-friendly, empathetic and professional.
Digital Mental Health + AI technology industry is developing fast around the world.
In Sophie Park’s 2019 research: “From Treatment to Prevention: the Evolution of Digital Healthcare”, it indicates that digital mental health is to solve the pain points of transportation cost, illness shame then help more people. Major forms include :
Website for treatment
Virtual assistant applications
Virtual chatbot applications
Gamified treatment products, etc.
Also, in China, the “2022 Digital Mental Health Services Industry Research Report” indicates that companies in this field are mainly divided into “online platforms” and “digital therapy products”.
After researching the industry, we decided to continue with consumer-level products that are targeted to the larger groups who are generally healthy but suffer from psychological issues. Realizing the strenghs of AI technology, we narrowed our focus down to AI chatbot therapist apps.
Competitor Analysis
We conducted the analysis below to better learn about the competitors of AI chatbot therapists:
The insights we learned from the competitor analysis are:
The global demand for AI chatbot therapist app is growing;
AI chatbot therapists are going to be more empathetic and personalized;
Users easily forget to get back, that is, the user stickiness is relatively low, due to the visual and interaction of the chatbot are not related to users’ personal experiences.
03 Primary Research
Questionnaire
To find the key needs & wants of AI chatbot app users, during the 26 days, we conducted two round questionnaire surveys with a total of 258 participants, then summarized responses in data visualization below:
Interview
Next, we interviewed 16 from the participants and summarized 4 user personas.
Then, we extracted key words from “Users’ Needs and Wants” and categorized them in to “Role“, “Space“ and “Talk“, which are characteristics of our AI chatbot product.
KANO Model Analysis
With the “Agile Development“ concept in mind, we sought to find the most important features of our product. So, we invited participants who expressed interest in our product to answer the 3rd round questionnaire then analyzed the results using KANO Model.
Fourth quadrant shows attractive features, while the first quadrant show the essential ones.
04 Define Product
At this stage, I drafted the Product Requirement Document as the product manager and here are the core contents.
Our product is a lightweight therapeutic app to help users find their “inner child” and get professional supports through conversation, has two key features :
therapy supported, relaxing conversations with AI chatbot ;
personalized roles and visuals of AI chatbot, generated by users.
User Flow
Product Stucture
05 Technical Support
Personalized Visuals of AI Chatbot
The initial proposal was to let users hand draw objects that are emotionally significant then make the stylized images move and have facial expressions through Animated Drawings tool.
Then our product was showcased in the International Consumer Electronics Expo 2024 in Shanghai, China and received feedback on this way to generate AI chatbot.
After the exhibition, we iterated our design in two aspects:
We continued to explore tools like ComfyUI, Animated Drawings and FaceMesh to realize our solution ideas.
Personalized Roles of AI Chatbot & Therapy Supported Conversations
By prompting in ComfyUI, we generated the AI chatbot with the personalized role which users want and can provide emotional counseling based on the “Inner Child Therapy“ mentioned before in Qijia Shi’s book.
Considering the visual effects of the user interface design phase later, we continued to prompt a conversation framework in the AI chatbot that can present a comparison of before & after using the AI chatbot therapy. This enables users to have a direct feeling of changes in moods.
06 UI/UX Design
Low-fi Prototype
With a clear mind of user needs & wants as well as what can be realized using AI technology, I worked with Hairu Zhai, the designer of our team, to draw the low-fidelity prototype :
Hi-fi Prototype
07 Future Outlook
End-to-End Multimodal Models
Powered by end-to-end multimodal models, more natural and efficient interactions can be realized to provide smarter solutions and innovative experiences for users.
Complex System Prompting
Large models will improve their understanding of prompting and execute tasks more accurately and smartly. As text commands replace more and more coding works, natural language interaction will simplify the development process and reduce costs.
As creators, we’d like to say ……
As AI technology continues to advance, many people harbor concerns about being replaced by cold, emotionless machines. However, our goal is to harness these technologies to support individuals with mental health challenges, combining children’s hand-drawn art and compassionate therapeutic conversations to provide warmth and connection. We believe technology should serve as a tool to enhance human lives, not replace us, and we envision a future where human-machine integration fosters greater well-being and harmony.
