Vera
Building trust into smart fitting rooms
Looped Video — Final Product
Role
Product Designer
Team
Solo
Timeline
Jul – Aug 2026
Context
Online shoppers guess on fit, fabric, and feel, driving return rates up to 40%. In-store, shoppers navigate crowded racks, carry armfuls of clothing, and wait for a fitting room – often having to repeat the process when an item isn’t quite right.
New tools are closing the gap between expectation and reality
Alvanon creates 3D avatars from shoppers’ measurements and photos to provide fit predictions.
Tommy Hilfiger partnered with Zero10 to introduce sales-floor AR mirrors for virtual try-ons.
Remaining Gaps
Privacy & Body Data
Body avatars and virtual try-ons rely on sensitive personal data, raising questions around how measurements and photos are collected, stored, and used.
Lack of Fitting Room Integration
Current tools stop short of the fitting room – a critical point where shoppers make final decisions and privacy matters most.
How might we bring AI- and AR-powered personalization into fitting rooms without compromising shopper trust?
This isn’t just an ethical choice – earning shopper trust is a competitive advantage, especially at the moment purchase decisions happen.
Research
Before designing anything, I interviewed 9 shoppers to understand their shopping habits and feelings toward tech in fitting rooms.
Key Insights
Transparency is non-negotiable
“I wouldn’t try [a fitting room with AI or AR] unless I understood how it worked…I know a lot of companies sell customer data.”
Thomas, 46
Privacy information can’t live in the fine print
“Obviously I want to know how my data’s being used, but no one’s reading those essays of terms and conditions.”
Amir, 27
Item-swapping breaks buying momentum
“If the size is wrong and nobody’s outside to help, I just leave the shirt behind.”
Maya, 19
Pseudonyms used for privacy.
Experience Flow
To address these friction points, I designed Vera (derived from the Latin veritas, meaning truth) – a smart fitting room mirror that builds on existing sales-floor AR.
Simplified flow diagram, with Insights, Preview & Find Your Size highlighted
Solution & Key Decisions
I designed the screens in Figma and used Claude Code for rapid prototyping, using several rounds of iteration to ground the final experience in 4 key decisions:
Defining the mirror’s boundaries
I initially explored a more feature-heavy concept where Vera handled password sign-in and checkout for maximum convenience.
As I considered the types of information the mirror truly needed, I realized these interactions required additional data collection without meaningfully improving the core fitting room experience.
I replaced password entry with QR sign-in and moved payment off the glass entirely, preserving human interaction for checkout.
Video of Sign In & Checkout Screens
Giving users control where it matters
Privacy controls are often buried in account settings, disconnected from the moment they’re actually relevant.
When a shopper requests a size recommendation, that is the moment they need to understand what information is being used, what is stored, and have the option to clear it.
This same principle applies to live AR previews.
Video of Getting Size Recommendation / Preview Interaction
Providing clarity without surveillance
Feedback intended to build transparency can backfire if it creates the impression of being watched.
An early version of the AR preview screen included a pulsing circle to signal that the system was actively rendering the item. It seemed like the obvious way to show “this is happening now,” but visually, it created the false impression of recording.
I removed the indicator and let the AR rendering communicate system activity itself.
Before — Pulsing Recording Indicator
After — Indicator Removed
Before
After
Making AI-driven insights feel like guidance
A recommendation stated as fact implies the machine knows more about a shopper’s preferences than they do.
In the Insights panel, I adjusted the language to use probabilistic terms like “often,” “frequently,” and “tends to” rather than definitive statements.
The difference is small on the screen but significant in the relationship it creates: one positions AI as an authority, while the other frames it as an assistant offering a starting point shoppers can accept, edit, or delete.
Video or Image of Insight Screen
Reflection
Designing for a mirror introduced technical constraints – like maintaining contrast on a reflective surface and keeping touch zones accessible.
But the hardest challenge was the precision required in communicating system intent and behaviour. A single word, icon, or animation can be the difference between a system that feels helpful and one that feels invasive, even when the underlying technology is identical.
I’m excited to bring this prototype back to my research participants for usability testing.
Image of Mockup