Reimagining Jewelry Creation through AI-Assisted Design
Founding Design, 2024
Jewelry design remains a highly traditional industry where creating new pieces often requires specialized CAD skills and close collaboration with manufacturers.
Skyveil explores an AI-assisted approach to jewelry creation, enabling users to transform simple descriptions or inspiration images into multiple design concepts and production-ready models.
My role
-
Designed the AI-assisted design workflow, enabling user to generate, refine, and iterate jewelry concepts
-
Collaborated with the founding team to explore product direction, including the transition from a consumer idea to professional workflows
-
Conducted industry research on traditional jewelry platforms and pricing structures to inform product positioning
-
Worked closely with AI modeling specialists to integrate generative design capabilities into the product experience
Outcome
-
Launched first product version and onboarded early B-side users in the jewelry industry
-
Early feedback showed strong interest from manufacturers and designers using AI for rapid design ideation and visual prototyping
-
Insights from these collaborations continue to inform future iterations of the platform
Scope
-
AI-assisted design generation
-
AI to production workflow
-
Early validation with industry users
Industry Context
some examples of jewelry CAD model
Jewelry design is typically a highly specialized process requiring professional CAD skills and collaboration with manufacturers.
As a result, turning a personal idea into a manufacturable design remains difficult for most people
Raises an important opportunity
Can AI make jewelry creation more accessible to everyone?
Original Vision
The original vision behind Skyveil was to enable anyone to design their own jewelry through simple descriptions.
Instead of requiring technical design tools, users could describe what they wanted using natural language or inspiration images. The system would then generate design concepts and transform them into structured jewelry models.
In this vision, a personal idea — whether symbolic, emotional, or aesthetic — could move directly from imagination to a design that could eventually be produced by a manufacturer.
The original web design
Product Direction Shift
Strategy — B2B First, B2C Next
The original vision was a consumer product: anyone could describe a piece of jewelry in natural language and watch it become a design they could take to a manufacturer. Beautiful idea. The early prototype proved the generation worked — but it also surfaced the real constraint.
The gap that defined the strategy
AI-generated concepts were creative and visually rich. But producing them required a second translation step: a trained designer or manufacturer still had to rebuild most of the model before anything could be cast. The bottleneck wasn't generation. It was the bridge between generation and production.
That meant a pure B2C product would either ship beautiful dead-ends or require a level of AI fidelity the technology wasn't yet ready to deliver.
The decision: B2B first, on purpose
Instead of forcing the consumer version before the engine could carry it, we sequenced the rollout. B2B first — landing the generation engine inside the workflows of designers and manufacturers, who could absorb the translation step as part of their existing process and tell us exactly where the AI needed to be sharper. Then B2C — once the engine was solid enough to bridge generation and production without a professional in the loop.
What the B2B phase looked like
-
Structured iteration controls (refine, regenerate variants, lock parts) aimed at trained designers
-
Onboarding flows reframed around design studios and manufacturers
-
Collaboration features so manufacturers and designers could share refinements across rounds
-
Tight feedback loops with professional users informing every model update
Two vision brochures of the different exhibitions
Outcome of the B2B phase
Skyveil shipped, onboarded early professional users, and validated the engine in real production contexts. Each round of feedback
sharpened both the AI output and our understanding of where the manufacturability bar actually sits.
Now: designing the B2C app
With the engine proven and the manufacturability gap closing, we're returning to the original consumer vision. I'm currently designing the B2C mobile app — a simplified experience that lets non-professional users describe jewelry ideas in natural language and generate personalized designs without needing technical design knowledge.
Same engine, redesigned around how consumers think and decide.
AI-Assisted Design Workflow
Skyveil explores how generative AI can bridge the gap between creative ideas and production-ready jewelry designs.
For designers and manufacturers, this significantly accelerates early-stage ideation and visual modeling.
Future Steps
Following early validation with professional users, the next stage of Skyveil explores a consumer-facing product.
The goal is to develop a simplified mobile experience that allows users to describe jewelry ideas in natural language and generate personalized designs without requiring technical design knowledge.
