AI-Driven Mac Desktop
Phase II

Reimagining the Mac desktop as an intelligent workspace through AI-driven creation and interaction.

Role AI Prototyping,
Product Designer
Timeline 2024 - present
(Part-time)
Company Cosmo Inc.
Team 2 Developers,
2 marketers, Ella

Overview

In Phase I, we focused on creating an intuitive platform to manage diverse personalization resources. As we researched the broader desktop experience, we found that the challenge goes beyond customization.

The Problem

01. Technical Barriers

Many customization enthusiasts have strong ideas but lack the coding skills needed to build widgets. Framworks, project structures, and deployment workflows make creation inaccessible to non-technical users.

02. Intent-Blind Interface Complexity

Traditional UI is intent-blind, exposing too many options and forcing users to navigate fragmented apps, tabs, and desktop surfaces just to complete simple tasks.

Opportunities

Analyzing these problems led us to define two major strategic opportunities for the next evolution of the desktop:

1

Non-Technical Creation

Enable users to create their own desktop features without needing to code.

2

Intent-Driven UI Widgets

Generate task-specific interfaces through AI integration based on user intent.

Design Insight

As AI improves in understanding natural language, reasoning through tasks, and handling execution, people no longer need to translate their intent into technical systems or fragmented workflows. AI can instead act as the bridge between what users want and how the system carries it out.

Future Blueprint

Transition from a simple customization tool to an AI-driven, intelligent, and adaptive desktop ecosystem.

Part 1

Lowering the Barrier to Creation

To address the first problem, we created the AI Widget Creation Guide.

Instead of requiring users to understand the underlying tech stack during the widget creation process, we provide structured context directly to AI. Users can simply describe their ideas in natural language, and the AI handles:

  • Intent interpretation from natural language
  • Widget layout generation with system styling
  • Technical implementation details
  • Deployment to the live desktop environment
AI Widget Creation Guide Framework

This shifts widget creation from a coding task into a communication task. Users no longer need to think in terms of files; they only need to express what they want to make.

How It Works in Practice

"Create a minimalist clock widget with a pixel cat that I can interact with as it walks."

AI Workspace Generation
Part 2

AI Desktop Agent

How could the desktop become a more intelligent interface for getting things done, instead of just a place where complex interfaces accumulate?

To explore this, we introduced a desktop-level AI agent approach through Cosmo Commands.

Cosmo Commands Beta

Instead of navigating through layers of apps, users can simply describe what they want to do. Cosmo then handles the steps and returns the result as a clean widget directly on the desktop.

✨ AI Automation

Automates navigation through browser sessions and returns exactamente what you need as a clean widget UI.

🤖 LLM Integration

Chat with ChatGPT, Claude, or Gemini directly from the desktop without constantly switching between tabs and apps.

check beta prototype

Where We’re Headed

From

Static destinations

Information-compact interfaces that reflect application boundaries.

To

Intent-driven interaction

Task-specific surfaces that emerge and dissolve as needed.

Cosmo is our attempt to explore that future: a workspace where intent becomes the interface.

Reflection

This project helped clarify that the future of desktop experience is not just visual customization. It is about reducing friction between what users want and what the system requires.