AI try on and wardrobe ecosystem
GOAL:
TIMELINE:
TEAM:
add a layer of personalization and possibility to a long-time essential for inspiration and dreaming. I designed an AI look generation and wardrobe ecosystem to strengthen Pinterest's commerce revenue.
1 week
me, because I love Pinterest & I wanted to show you a passion project
EXPLORATION
Pinterest has something unreplicable: influence, aspiration, and 560M monthly users who trust the platform to understand them.
However, Pinterest's business model depends on shopping revenue. Pinterest reported 600M users and $4B revenue in Q3 2025—but notably, no shopping metrics. This suggests their shoppable Pins underperform despite 270M users engaging with fashion monthly.
Pinterest hasn't fully captured the commerce opportunity. Despite introducing shoppable product tags, the feature significantly underperforms relative to engagement.
Natural, low pressure space to build your wardrobe and shop on Pinterest.
AI Stylist
AI Try On
Virtual Closet
A conversational partner that curates complete outfits for real contexts, like a wedding, weekend trip, or coffee chat.
A digital self that tries outfits for you, making it easier to see how clothes might fit your style and body.
A living wardrobe sourced from the Stylist's generated looks. Everything is shoppable, nothing requires purchase.
Strategic Value:
Positions Pinterest in $60B AI fashion market
Creates defensible moat (user built closets = switching costs)
Projected $40M revenue impact from 20% of users in 1 year.
For Users:
Organization: Saved Pins become actionable wardrobes
Confidence: Try on removes "will this look good?" anxiety
Utility: Personal styling tool, not just inspiration board
THE RESEARCH
WHY SHOPPING DOESN'T WORK ON PINTEREST TODAY
I spent a week analyzing Pinterest's current fashion and shopping experience to understand where it breaks down. Here's what I found:
WHAT PINTEREST LOOKS LIKE NOW
Pinterest's Q3 2025 earnings show:
Pinterest introduced shoppable product tags to monetize fashion content
~270M users engage with fashion Pins monthly (45% of total MAUs)
$4.06B annual revenue
No reported conversion rates, shopping engagement, or commerce-specific revenue
What this suggests: Investors and Business insight newsletters, including Forbes, suggest that shopping is underperforming relative to its potential.
THE CURRENT EXPERIENCE
Shoppable product tags are underutilized:
Users save fashion Pins as inspiration
Some Pins have tagged items with "Shop" links
Shop links often don't match the image
Users don't click through at meaningful rates
No visualization tool (can't see items on themselves)
No styling guidance
The Behavior Pattern:
Users browse Pinterest for fashion inspiration → Save 100+ outfit Pins → When they actually need to shop → Look at Pinterest for inspiration → Close the app and search for the actual pieces elsewhere.
Pinterest captures inspiration. Competitors capture commerce.
THE INSIGHT
To understand the user's perspective, I analyzed Pinterest user forums, fashion subreddits, and Twitter conversations. I also spoke with 5 friends who are heavy Pinterest users to understand their shopping behavior.
Key insights from users:
USER PERSONAS & NEEDS
Sara | 23 | Recent Grad
"I have a networking event next week and no idea what to wear. Models and influencers never look like me."
Jasmine | 29 | Fashion Designer
"I've saved so many outfits, but scrolling through them just stresses me out. I never end up buying anything even if I want to."
Michael | 24 | Medical Scrub
"I need to update my wardrobe. Everything I like is saved on Pinterest, but using it to shop feels like I'm doing all the work."
Needs:
Organize saved items by occasion (work, events, casual)
See outfits styled on her body type, personalizable avatars
Needs:
Direct links to shop both complete looks and individual pieces
Dedicated place to view personal wardrobe and saved pieces
Needs:
AI guidance on how to complete outfits and why pieces work together
Inclusive male styling (most fashion tech is women focused)
INDUSTRY CONTEXT: WHY NOW?
Fashion Tech is Exploding:
Fashion Tech market: $240B (2024) → $345B (2030) — 44% growth
AI in Fashion: $2.2B (2024) → $60B (2034) — 2,600% growth
Virtual try-on adoption: +30% year-over-year
Competitors Are Moving:
Amazon: Building AR try on (25% conversion lift, 30% fewer returns)
Instagram: Pushing creator commerce and AR shopping filters
Stitch Fix: AI personal styling (but no discovery layer)
ShopMy: Creator commerce with curated items to shop
Consumer Behavior Shift:
Post COVID: People shop more online, yet how an item will look on you is a guessing experience.
Gen Z expectation: someone should have figured out how to shop better
Rising interest in sustainable fashion: People want to buy confidently to reduce returns
The opportunity: Pinterest owns aspiration first discovery. 270M users curate their ideal style over months and years, massive intent data currently sitting. Pinterest can differentiate by focusing on aspiration + styling guidance, but keep its low pressure edge.
DESIGN IMPLICATIONS
Make shopping feel natural, not forced
Pinterest is low pressure intentionally
Don't interrupt browsing with "BUY NOW" prompts
Let users build closets at their own pace
Shopping happens when they're ready
Build on existing behavior
Leverage Pins users already save
Don't require learning new patterns
Feels like Pinterest, not a separate app
Create persistence
Closet accumulates over time (not ephemeral)
Users invest in building it
Becomes a daily tool, not one-time novelty
Ensure inclusivity
Diverse avatar options (all body types, skin tones)
Gender-neutral design
Accessible to all users
Respect privacy
Opt-in avatar creation
Clear data usage policies
Easy delete options
PRODUCT GOALS
USER FLOW
KEY FEATURES
PROFILE HUB
For the User
For Pinterest
Immediate access when they need help:
Persistent access to their wardrobe:
Choice based on context:
Two entry points match two different needs:
"I want new ideas and looks" → Tap Stylist button
"Let me browse what I have and shop" → Tap Closet tab
Feels native, not foreign:
By matching Pinterest's existing design language (red pill button, tab structure), Virtual Stylist feels like it's always been part of Pinterest, not a confusing new tool to learn.
This is the first tab explicitly tied to shopping/action (not just inspiration).
Prominent placement (profile level button + third tab) means users will actually find the feature. Hidden features don't get used; visible features create engagement.
Both paths lead to shopping, but serve different user states.
Creates new Pinterest behavior:
Home tab = discover content
Saved tab = save inspiration
Closet tab = organize wardrobe and shop
STYLIST PROMPTING
For the User
For Pinterest
Immediate help without setup:
No onboarding, no questionnaires, no configuring preferences. Just ask for what you need and get styled. The fastest path to value.
Speak naturally or use structure:
Flexibility allows users to choose their comfort level.
AI does the creative work:
Users don't need to know how to style outfits or what pieces work together, unless they want to. This removes the intimidation of having to take pictures.
Feels like a conversation, not a form
Low barrier allows for high adoption:
Familiarity and less time spent figuring out the feature increases the likelihood someone actually uses it.
Captures rich intent data:
User prompts reveal exactly what they're shopping for - gold for serving relevant product recommendations and ads.
Feels helpful, not transactional:
Unlike "SHOP NOW" prompts that feel pushy, conversational styling feels natural and helpful, aligned with Pinterest's values and voice.
CURATED LOOK SCREENS
For the User
For Pinterest
Transparency + ability to shop:
Every item is labeled and shoppable. No mystery pieces. Users know exactly what they're looking at and where to get it.
Flat lay format allows the user to visualize the look's details before seeing it on their avatar, where layers and accessories aren't as visible.
Iteration is encouraged, not penalized:
"Edit" option signals that AI's first attempt isn't final and treats the Stylist as a collaborator. Mirrors real life, set the outfit, then wear it.
Outfit based shopping drives larger transactions:
Industry data: Outfit based shopping converts 2-3× higher than single items and increases basket size (users buy 3-5 items vs. 1). Showing complete looks directly impacts revenue.
Guiding users to visualize before saving creates a qualification step. More time invested and higherintent = higher conversion.
TRY ON SCREENS
For the User
For Pinterest
Like a fitting room. You can assess the look comprehensively, whether the details fit each other.
Your inspiration becomes personal. Seeing yourself in an outfit you saved that fit an inspirational life you imagined, is now personal to you, it feels attainable.
Industry data: Virtual try on increases conversion by 25% (Amazon's published results). Users who visualize before buying are dramatically more likely to purchase.
There's a psychological phenomenon called the endowment effect: once you see yourself wearing something, you feel ownership. Try on creates emotional attachment before buying, increasing likelihood users complete the purchase.
VIRTUAL CLOSET
For the User
For Pinterest
Has 2 main functions:
Serve as a visual, direct shopping list that's fully personalized to you.
A virtual wardrobe for users to plan their own looks based on Pins, and pieces they've actually bought from Pins.
Pinterest becomes more of a daily tool, when users are browsing their own closets for inspiration on Pinterest.
Every closet item is shoppable, you're constantly exposed to the opportunity to realize your closet, though it's never pushed.
How users organize folders, which outfits they save, how often they browse all signal shopping intent. This behavioral data improves recommendations and Pinterest's algorithm.
IMPACT
Year
Users
Incremental Value
1
54M (20% adoption)
+$40M
2
108M (40% adoption)
+$108M
3
162M (60% adoption)
+$198M
REFLECTION
The best features feel like they were always missing.
This was the last step that validated my design. I showed my sister my design, and she was so excited, she went to update Pinterest to try it out.
Established products have invisible rules you have to learn. Mission and voice matter
Pinterest has been trusted and use for inspiration for so long. It's designed for stability, with clear inspiration and personalization mission, which is something to maintain as priority as I design. I did this as a passion project because I love Pinterest, so why would I change it?
Introduce as little new behavior as possible.
Users already know how to use Pinterest, and appreciate its simplicity. Virtual Stylist leverages those patterns so the only truly new thing is talking to AI. The closet tab and shopping feel like a natural next step.