LIU CHANG
04 / AI PRODUCT CASE STUDY

Desktop Pet

A rapid validation experiment spanning user need, a macOS prototype and two social posts with about 570,000 views.

  • Product Planning
  • AI-assisted Build
  • Electron

Context

Starting from desktop companionship and lightweight productivity, explored a desktop product combining a character companion, to-do reminders, hydration reminders and customization.

Product Build

Independently defined the core interactions and product rules, then used Codex and Electron to build a macOS prototype with a transparent always-on-top window, character dragging and locking, click-through behavior and continuous head tracking.

Validation

Iterated on animation anchors, motion interpolation, tracking smoothness and desktop interruption, then validated demand through two social posts with approximately 570,000+ views and 25,000+ likes.

RoleProduct Planning / AI Workflow / Prototyping / AI-assisted Development
StageInteractive macOS prototype
DP

Desktop Pet

57万+Views across two demo posts
2.5万+Likes across two demo posts
EvidenceSocial demand validation / Two demo posts
01 / CONTEXT

context

Starting from desktop companionship and lightweight productivity, the project explores a product that combines a character companion, to-do reminders, hydration reminders and customization.

Make utility feel lighter and companionship feel natural on the desktop.
02 / PROBLEM

problem

Desktop reminder tools can feel rigid, while companion-only characters may lack lasting utility. The product must balance function, emotion and desktop interruption.

Companionship should not interrupt work, and utility should not break the character experience.
03 / INSIGHT

insight

Desktop companionship works best when it is low-interruption, persistently visible and lightly interactive. Reminders and customization can create reasons to return.

The character is the entry point; lightweight productivity creates lasting value.
04 / SOLUTION

solution

Designed a desktop-pet prototype that brings character interaction, to-do reminders, hydration reminders and customization into one lightweight experience.

Use companionship to carry reminders and give productivity features emotional value.
05 / PRODUCT DESIGN

product design

Independently defined core interactions and product rules, including character drag and lock, click-through behavior and continuous head tracking toward the cursor.

Interactions need to feel alive without getting in the way of desktop work.
06 / AI WORKFLOW

ai workflow

Used Codex and Electron to build a macOS prototype, translating product rules into testable desktop interactions and refining motion logic throughout development.

AI-assisted development shortened the path from interaction concept to testable prototype.
07 / BUILD

build

The prototype implements a transparent always-on-top window, character dragging and locking, mouse click-through, and continuous head tracking.

The concept has been advanced into an interactive macOS prototype.
08 / VALIDATION

validation

Iterated on animation anchors, motion interpolation, tracking smoothness and desktop interruption, then tested audience interest through two social-media product demos.

Technical smoothness and perceived interruption directly affect whether users want to keep the character on their desktop.
09 / RESULTS

results

Two product-demo posts received approximately 570,000+ views and 25,000+ likes, providing an early demand signal for desktop companionship and lightweight reminders.

Social data validates concept appeal, but not long-term retention.
Two Desktop Pet social demo posts showing 64,000 and 512,000 views
Demand signal: approximately 570,000+ views and 25,000+ likes across two demo posts.
57万+Views across two demo posts
2.5万+Likes across two demo posts
10 / REFLECTION

reflection

Social response demonstrates the concept’s appeal. The next step is to validate real usage of reminders, long-term retention and whether customization creates sustained motivation.

Moving from “interesting to watch” to “worth using long term” requires in-product validation.