Dongpei Yue is
a service & experience designer bridging AI, gameful interaction, and human relationships, with a foundation in architecture and spatial thinking.

Product & Service Designer
Architect

Photographer


About me
My Blog (In Progress)





Alter Emo
Category:AI Emotional Well-being/
Skills:User Research,AI-Driven Strategy,System & Service Design,Project Management,UX/UI Design.
Roles:Service Designer,User Researcher,Experience Designer,AI Strategist,Prototype Developer/
Duration:6 Months/


“In a world where efficiency is automated, what remains profoundly human is how we feel—and who understands that.”




In the current landscape of AI-powered emotional support tools, young professionals aged 25–35 represent the core user group. Facing high emotional stress, loneliness, and career uncertainty, they actively seek solutions for emotional regulation, stress relief, companionship, and personal growth.
Products in this space tend to deliver value through two primary dimensions:
  1. Instant emotional relief: Offering quick and accessible support through AI conversations, guided meditations, and auditory relaxation techniques during moments of distress;
  2. Co-growth and emotional resilience building: Fostering long-term engagement through companion-like relationships, emotion journaling, and psychological training, enabling users to better understand themselves and grow over time.

Not Sick, Just Stuck“I used to stress about whether I was doing a good enough job. But lately I’ve been thinking—does this job even mean anything to me?”
We conducted interviews with 25 young professionals across various industries to explore the emotional triggers they experience in the early stages of their careers and how they cope with them. 

What emerged was a notable shift: while performance anxiety remains present, many are increasingly turning their attention to deeper reflections on the meaning of work—questioning the true value their jobs bring to their lives.
General Emotion Coping Patterns
In addition, we organised the different coping strategies into a pattern to better understand the challenges at each step. This pattern serves as a foundation for informing our design decisions
Problem Statement “Evenings (approximately 8 PM–11 PM) are a critical emotional vulnerability window for young professionals—a largely unserved mental health support gap.” Calm Workplace Report (2024); Young professionals often face uncertainty about career identity and direction. External triggers like social media and peer comparisons heighten the pressure for quick clarity, clashing with the complex nature of these decisions. With emotional resilience still developing, they’re especially prone to self-doubt and anxiety—most acutely during solitary evening hours.


Coping Mechanism & Usage of AI
Second Round Workshop Materials
Second Round Workshop Materials

We held a second workshop to gain insights into the coping mechanisms about solutions participants relying on. and the usage of AI. 
Through workshops, we identified two common pathways for emotional processing among young professionals:
  1. Regular reflection with trusted people helps them gain both empathy and practical feedback.
  2. Externalizing thoughts—through writing, talking, or recording—allows them to clarify emotions and organize internal chaos.

Even after we know the solution (e.g. reflection, expression), users still face real barriers to doing it While reflection and emotional externalization are commonly effective,
users still face real-life barriers when trying to adopt these methods.
This leads us to examine the micro-level challenges users encounter during use.

What Gets in the Way — Even When Users Know What Works
Second Round Workshop Phot

While users share common emotional struggles at night, their coping styles and trust in AI vary greatly.This creates a complex design space where a one-size-fits-all solution falls short.

What they need is a system that feels structured but emotionally safe, one that adapts to their shifting emotional states, and respects their personal distance with AI.

Based on these needs, we propose the following hypothesis to guide our design and validation process.


Hypothesis We believe that a context-aware and emotionally adaptive AI care system can provide meaningful support for young professionals who face emotional vulnerability and identity anxiety at night.

Ideation
By combining CBT-based journaling, therapeutic roleplay, and emotionally intelligent companions, the system can create a safe, reflective, and engaging space to process self-doubt and rebuild inner clarity.

We aim to validate whether AI—when designed with empathy, adaptability, and ritual—can effectively personalize emotional coping support in moments of quiet struggle.

“Not by replacing human connection, but by extending it when it’s most needed.”


Serious Play:Light Relief, Real Healing

Agent Prototype
“You are not alone — a softer version of you is always listening, growing gently beside you.”
In a small town lives your digital self — a character shaped by your behaviors, attitudes, and emotions. This digital you inhabits a rich, evolving world filled with everyday happenings, close friends, and kind-hearted neighbors. Through your phone, you stay connected with this version of yourself, sharing stories from both your worlds. As time goes on, you listen, support each other, and grow together.
Introduction:
Alter Emo is a pocket-sized emotional companion AI, brought to life as your personal avatar in a vibrant digital town.
In moments of emotional challenge, it offers gentle reassurance and structured guidance. Over time, it helps you build the habit of journaling, review your growth curve, and strengthen your emotional resilience.
Its ultimate goal is to help you stay true to your own direction — free from the sway of short-term external 
pressures or emotional turbulence.
Non-Linear problem solving:
In addition to conversations with the Digital Self, users can also choose to talk with different residents of the virtual town. Each resident comes with their own background and personal story, offering unique perspectives to help users reflect on their thoughts and navigate their questions.
Interview with Leah (Incoming PhD candidate in Computational Media at the University of California, Santa Cruz,specializing in serious game and service design
High Engagement & safe expression: Look and FeelWe adopted a pixel-style game aesthetic, drawing inspiration from the nostalgic games many young professionals played during childhood. By echoing familiar visual forms, we aim to create a comforting and emotionally safe space for users — one that feels playful, approachable, and open for honest dialogue.


Empower and Reflection:
We also designed a Quick Journaling feature as a fast-entry point for emotional expression. Users can select emotion keywords, tags, and specific events to quickly capture how they feel in the moment.

Tangible Thinking:
To encourage users to consistently engage in emotional journaling, we designed a subtle feature: the Digital Self gradually adds symbolic decorations to the user’s virtual “home” based on their past conversations. Over time, as journaling continues, each user’s space evolves to reflect their unique emotional journey, turning their commitment into a personalized and visualized sense of progress.
AI Companion - Your Digital Self
AI Interview Interface from“Generative Agent Simulations of 1,000 People”
“Imagine an AI that learns from your attitudes and behaviors—by observing and talking to it, you gain new perspectives on your emotions.”

By inspiring from the paper “Generative Agent Simulations of 1,000 People” we see the promise of human behavioural simulation. General - purpose computational agents replicates human behaviour across domains brings highlight for our project.   

“Generative Agent Simulations of 1,000 People
Workshop Summary
We conducted a 1:1 interview-style workshop to ensure user privacy. The aim was to test how effectively AI could gain a basic understanding of the user’s behaviors and attitudes within a 15-minute conversation, as well as to evaluate whether the AI’s tone of response made users feel comfortable.
Findings1. Users preferred the AI’s responses during the initial onboarding stage.

2. Based on the original research, we reduced the personality and behavior learning process from 2 hours to 15 minutes.Accuracy remained at 75%, despite the significantly shorter learning time.


CBT - Journaling

“Every conversation is a small step inward — toward clarity, healing, and the quiet understanding of yourself.”
After completing a few simple steps — such as talking to your Digital Self, doing a Quick Journaling session, or chatting with town residents — users typically spend around 20 minutes. At the end of this journey, a Daily Emotion Journal is automatically generated, summarizing the day’s reflections.

All conversations are gently guided by principles from Cognitive Behavioral Therapy (CBT), helping users recognize thought patterns, explore emotions, and understand the connections between their thoughts, feelings, and behaviors. The dialogue flow is structured as follows:

Validating Our Concept With Our Audiance(Exhibition)

Exhibition ObjectiveThe primary goal of this exhibition design was to test audience interest in both the look and feel of the concept and the logic behind AI-generated emotional journaling. Unlike traditional handwritten or private self-recording methods, our approach encourages users to record their emotions through AI-driven games and scenario-based conversations. The exhibition also aimed to gather user feedback for further refinement.
Print for daily emotional Diary