Unlocking the Possibilities
of Old Messages

DURATION

4 months

Jan - Apr 2026

TYPE

Thesis Project at

ArtCenter College of Design

MY ROLE

Solo Project

I focused on feature discovery by exploring AI text generation through coding and UX/UI design.

TOOLS

Figma, After Effects, HTML, CSS, Javascript

AI Search to bring users joy and emotional support
through memories in old messages.

1

Consent to Data Collection & Privacy

We use your message data to power AI
search and recommendations—only with
your consent.

01

2

Ask, show me a funny story from college!

Users can access it directly from the search bar. For example, searching “Any fun stories from college?” surfaces relevant messages from past conversations.

Working Prototype with OpenAI API

I created a working prototype to explore how the product would work in real use and
to iterate on the project through usability testing.

Behind Scenes of Design

I added around 40 messages to the chatbot and asked it to select the most relevant one based on the user’s question.

What I've learned

  1. Depth of Problem Definition

At first, the problem was defined broadly as Gen Z not fully utilizing bookmark data. Because it was too broad, the solutions were vague and

  1. Understanding the Ecosystem

Initially, I assumed this was mainly a problem for content consumers, so I only interviewed consumers. However, when designing, I realized the

  1. Importance of Iteration

By creating multiple versions (A, B, C, and even D), I experienced how iterative design makes solutions more practical and user-friendly.