Ji Kim
Visual, Interaction, Experience Design
Currently open to new projects
2021~2024
Cartier Retail Innovation Lab
Role
UX/UI, Interaction Design, Research, Visual Design
Overview
Looking Glass is an in-store virtual ring try-on experience designed for Cartier. It allows clients to visualize unavailable rings on their hands, guided by a trained sales associate. I led UX/UI and interaction design across research, prototyping, and scaling, shaping an interaction model that feels intuitive, invisible, and true to Cartier’s high-touch brand.
“How might we design an AR try-on experience that feels so natural, clients forget the ring isn’t real—without ever touching the technology themselves?”
By the time I joined the project, several foundational decisions had already been made
- Cartier had committed to building an iOS-based AR app with a custom-designed table lamp containing a camera.
- The team had tested other try-on methods (VR headset, smartphone AR, magic mirror) and ruled them out because they did not meet Cartier’s standards for quality or hospitality.
- The team chose GAN AI-generated ring images instead of 3D models to achieve higher realism in the product visuals, especially for reflections, metal, and stones.
- The AR imagery should be real-time video so clients can forget that the rings are virtual.
My Role
My role began as the team moved from the early hardware prototyping to defining the user experience. I led UX and interaction design across the full lifecycle from early research to pilot rollout.
Responsibilities Included
- Defining user roles, flows, and behaviors
- Designing UX, UI, and interaction patterns
- Creating wireframes and prototypes
- Aligning visual design with Cartier’s physical environment
- Supporting engineering, QA, and training content
- Planning and running user testings
Cartier boutiques often don’t have the exact ring a client wants due to rarity and intricate sizing.
This Created Friction
- Clients couldn’t try on what they came to see, and the brand lost sales
- Sales relied on product photos or showing similar products
- The brand lost money on product transfers and size exchanges
Business Goals
- Visualize the unavailable rings in accurate scale and materials
- Maintain trust and confidence in the moment of purchase
02. Early Findings: MVP and User Testing
All core UX findings were shaped early, between MVP development and moderated user testing.
We Tested Assumptions
- Should the client control the experience?
- Would a self-service iOS app feel empowering—or uncomfortable?
- How does a digital screen interact with Cartier’s signature sales flow?
What We Learned
- Clients didn’t want to operate the app on their own
- Sales associates were central to maintaining the luxury feel
- The more we minimized UI, the more immersive the experience became
This reframed the project. We shifted focus from client-led UX to sales associate-guided interaction design.
03. Designing Nonintrusive Interaction
UI as Mirror
The core experience was visualizing a ring on the client's hand in real-time with fewer interruptions. That meant the UI shouldn’t be overpowering. As a result, the Looking Glass interface feels luxurious and imbued in the space.
Key Design Decisions
- A camera feed calibrated to life-size hand scale on the screen
- Hidden interface elements that only appeared when actively needed
- UI and its transition that mimicked real-world Cartier sales tools (ie. Tray moves in and out)
- Visual styles based on Cartier’s interior finishes—not their e-comm style guideline
CTA displayed in all states. Most users answered it feels complicated and inconsiderate
Mirror State: minimal to no UI shown
Active State: CTA moves in, reminding the tech Hand size on the screen calibrated to life-size
Ring Swap Interaction
During a sales appointment, sales associates frequently needed to swap out the three rings in the “Tray” with other products in the “Catalog” page when clients wanted to try multiple rings. We tested three interaction patterns to determine the most effective interaction to accomplish this task.
The winning version (v.1.5) reflected sales associates’ typical behavior in boutiques—quick, light, and tactile.
Task
Swap a ring between the Catalog and the Tray. Enter the active “ring swap” state.
- v.1.3: Long-press a ring image (Low discoverability)
- v.1.4: Same as above, but the pop-up instructions added (Limited impact)
- v.1.5: Double-tap a ring image (Most intuitive)
Long-press to enter the active state. Benchmarked the iPhone OS pattern to reorganize the home screen icons. Most users frequently failed to discover the feature.
Pop-up instructions added due to the frequent task failure in the previous version. However, the users (SA) still failed to accomplish the task because they don’t have time to read the instruction while helping clients.
Double-tab a ring image to enter the active state. Users discoverability was higher than long-press. Added shade to highlight the active items and more direct GUI to help the user to accomplish the task.
04. Beyond
The experience was mainly designed to enhance sales at Cartier boutiques, but we prototyped other use cases beyond that. By updating a few interaction patterns on the tray and the catalog page, the experience can be operated by less experienced staff or even by clients. We found great potential to explore off-site and non-sales usages with small to medium catalogs.
05. Outcome and Impact
Looking Glass was installed in flagship boutiques globally and supported successful sales of products that weren’t physically present. It provided a luxury AR experience that was both scalable and deeply on-brand.
- Created direct sales during the rollout
- Rolled out in 6 boutiques in North America and Tokyo event
- Boosted sales associate confidence in adapting new tools
- Informed future digital strategy for retailtainment experiences
- Helped define Cartier’s digital interaction language
06. UI Design System
Reflection
This project taught me how subtle interaction and visual design decisions shape trust, luxury, and emotional bonds with users when they are asked to adopt a new, challenging tool. What makes it successful is that it lets the users bridge the technical gaps with their own physical behavior, comfortable pace, and familiar elements instead of feeding them the common ideas of fast, quick technology solutions.