01 · The question
Moodboards feel personal. Online stores feel like billboards.
In early 2025, a product manager and I kept asking one question: why does a moodboard feel so personal? In creative work, it comes before anything else. It says: this is made for you.
E-commerce does the opposite. You land on a homepage and get banners, promotions and seasonal pushes. It’s the same store for everyone, with no room for the person actually there.
That gap became our concept: a store that keeps reshaping itself around one buyer. Not just recommendations. The buyer walks into their store.
02 · Research
Gen Z wants a store that knows them, and trusts people over brands.
We had 2 weeks to pitch a heritage denim retailer that was struggling to win and keep Gen Z in a crowded, increasingly AI-native market. Week one went to research: the retailer, its competitors, and what Gen Z expects from a store.
Taylor, 16
Persona · high-school student, US
Shops to express herself, but is overwhelmed by generic curation.
What Gen Z told us (survey)
- 84%expect a store customised to them
- 76%trust peers and creators over brand ads
- 92%are fluent in swipe interfaces
Where personalisation stood
Key insightPersonalising the whole experience, not just the product recommendations, was the opportunity nobody had taken.
How we researched
- SWOTThe retailer’s position
- PESTELForces shaping retail
- Porter’s five forcesCompetitive pressure
- Brand auditType, colour, imagery, layout
- Gen Z surveyExpectations and trust
| Force | Pressure |
|---|---|
| Rivalry between brands | High |
| Substitutes | High |
| Buyer power | High |
| New entrants | Medium |
| Supplier power | Lower* |
*Not rated in the original analysis; described as reduced by a large supplier base.
03 · The idea
Don’t rebuild the store. Give every shopper a wardrobe inside it.
Rather than reimagine the whole storefront, we proposed one feature inside the existing experience: My Wardrobe.
The metaphor was deliberate. A wardrobe is intimate. It’s yours, and every compartment knows what belongs in it. In My Wardrobe, every card is an agent that talks to the others, learns and curates, all the time.
- An entry point on the store’s own homepage
- Always one tap away in the tab bar
Patchwork
seasonShop the drop →
04 · The system
One buddy you teach, plus agents you create.
- Style Buddy runs the wardrobe. It’s the default agent at the centre. Every swipe, purchase and chat teaches it more about you.
- Shoppers make their own agents, each for a purpose. Taylor might create Combinator (outfits from what she owns and what’s new), Blue Jeans for Taylor (only denim that fits her taste) and Care Library (care and longevity).
- The agents talk to each other, so Taylor never has to manage them.
Try itAsk a question, open a pick or add it to your bag, then go back: the chat is saved as a card. Tap Edit to delete cards.
Combinator3 new outfitsBlue Jeans for Taylor5 picksCare Library2 reminders
✦ Fit Twin: S fits you best in this cut
- Style Buddy sits at the centre
- Agents Taylor made herself
- Agents sharing what they learn
05 · Features
Six features, each borrowing a habit Gen Z already has.
Patchwork dress$58
Two-tone jacket$74
Patchwork set$96
Swipe to teach
Borrows: swipe culture
Before browsing, shoppers swipe on up to 12 products. Each swipe trains the agents, using a gesture Gen Z already has muscle memory for.
Try itDrag the card, or tap ♥ / ✕.
Fit only. Your face is never shown or shared.
See yourself
Borrows: asking friends
AR shows the looks on the shopper, so Taylor doesn’t have to imagine the fit. Her photo never leaves her phone, and she can delete it any time.
Try itAsk friends to vote, or delete the photo.
How do friends vote?
- Taylor picks up to 4 looks and taps Ask friends.
- Tidewell makes a private link she drops into her group chat. No friend needs the app or an account.
- Friends vote in their browser. They see the fit only: renders are cropped at the shoulders, so her face and photo are never shared.
- Votes land live on her screen. The link and the renders expire after 24 hours.
How I style the wide-leg, 3 ways
Worn by 1.2k buyersReal people wearing it
Borrows: trusting peers
Every product page carries videos of creators and real buyers wearing the piece. Trust built the way Gen Z builds it: through people, not brands.
Try itPlay the creator clip.
Understood as
Search like you talk
Borrows: texting
Search understands full sentences and feelings, not just keywords. Taylor searches the way she thinks, and the store works out what she means.
Try itRun the search.
92% of shoppers with your fit kept the S.
Fit Twin
Replaces: ordering two sizes
Learns Taylor’s fit from what she kept and returned, each garment’s real measurements, and shoppers built like her. It recommends a size and says why. She still chooses.
Try itPick a different size.
- Patchwork midiSold out in S · watching
- Denim truckerWaiting for a price drop
- Denim midi skirtNew drop Fri, 10am
Drop Watch
Borrows: drop culture
Watches restocks, drops and price cuts in Taylor’s size, inside rules she sets. When something lands, it holds it for 20 minutes and asks. It never buys without her tap.
Try itSimulate a restock.
Why the last two need a store. The question I hear most is whether a chatbot could do this. It can talk about fashion, but it can’t see what Taylor kept or sent back, what’s in stock in her size, or act for her. Fit Twin and Drop Watch can, and she stays in the loop. We designed these two after the pitch.
| Capability | General chatbot | Tidewell agents |
|---|---|---|
| Your fit, from what you kept and returned | ✕ No | ✓ Yes |
| What’s already in your wardrobe | ✕ Only what you type | ✓ From your orders |
| Live stock in your size | ✕ No | ✓ Yes |
| Acting for you: hold an item, add to bag | ✕ No | ✓ With your OK |
06 · Outcome
The retailer became a Deloitte client. A multi-billion-dollar win.
The research, the concept, the system and the agentic interactions all came together in a 2-week pitch.
What I took awayThis project wasn’t born for the client. It started as a design question, and it won because we’d done the thinking before the brief arrived. That’s curiosity compounding: a concept from the margin of a notebook became the centre of the pitch.
