AI-Assisted Development

Changing what's worth building.

3D render of a lightning bolt, symbolizing fast AI-assisted development

Problem

Creator Hub is a desktop app used by creators to build content for Decentraland.

Over time, many small and large improvements were identified, but traditional development processes made some of them difficult to prioritize.

Opportunity

Use AI-assisted development to quickly prototype and implement improvements without replacing Product Thinking.

Goals

Deliver postponed quality of life improvements

Shorten implementation time

Validate a new development workflow

My Role

I worked as both Product Designer and rapid prototyper.

Workflow

Observe the problem

Define the UX solution

Prototype with Claude Code

Collaborate with engineers

Validate and deploy

Before and After

Multiple small improvements fundamentally changed how the editor feels to use.

Before

Creator Hub editor interface before AI-assisted improvements

After

Creator Hub editor interface after AI-assisted improvements

Examples

Snap and Grid Controls

Improvements: New contextual toolbar • More accessible snap and grid controls • New snap to object function • New snap to floor function • New Mathematical operations in the input fields • Fixed several input fields inconsistencies

Object snapping was initially estimated to take 6 weeks. A functional version was delivered in days.

Examples

Gizmos and Object Selection

Improvements: Object Hover and Selection appearance • Move, Rotate and Scale Gizmos • Gizmo colors and transparency • Gizmo hover and click zones

Gizmo interactions became more precise and predictable.

Examples

Scene Toolbar

Improvements: Tool button appareance and position • Multi-select support • Tool keyboard shortcuts • Visual feedback • Fixed tool enabled/disabled states • Fixed various bugs

Unnecessary friction was removed without requiring a large development effort.

What AI did NOT replaced

Product Thinking

Prioritization was still based on what we should build, not just what we could build.

User understanding

Design decisions were based on user feedback and our understanding of creators.

Design decisions

Interactions were designed first, then implemented with AI.

Engineering collaboration

Branches, pull requests, code reviews, and merges remained a collaborative process with the engineering team.

Learnings

AI can accelerate product development when applied to well defined problems.

Product Thinking should remain at the center of decision making.

Smaller improvements become easier to justify and deliver.

Rapid experimentation shortens feedback loops.

Thank you!