MyntraDigitalWardrobe
A concept extension to Myntra's My Stylist AI, letting users style their existing wardrobe alongside new purchases
In short
Myntra Digital Wardrobe is a concept feature I designed to extend Myntra's existing My Stylist AI. Right now, My Stylist only recommends outfits based on what you have bought on Myntra. This concept lets users catalog clothes they already own from anywhere, so recommendations reflect their real closet, plus a conversational AI stylist that replaces Myntra's current static style quiz. This is a self directed concept, grounded in real competitive research, not a shipped project.
The problem
People wear a closet, not a catalogue. Myntra's styling advice can only see the items bought on Myntra, so it recommends outfits around a small and often unrepresentative slice of what someone actually owns.
The result is advice that feels generic and always ends in a purchase, even when the right answer is already hanging in the wardrobe. Preferences are captured once through a static quiz and never revisited, so the app has no way to learn that someone is dressing for a wedding this week and a new office next month.
Myntra already holds purchase history, size data and a large catalogue. What is missing is a picture of the closet those purchases join, and a way to talk to the stylist rather than fill in a form.
What Myntra already has
The concept is built on top of what already ships today, not around it.
| Existing feature | What it does | Gap this concept addresses |
|---|---|---|
| My Stylist AI | Suggests outfits and pairings, but only from items bought on Myntra. | Most of what a person wears was never bought on Myntra, so the styling advice covers a sliver of the real closet. |
| Style Quiz / style profile | A one-time static questionnaire that sets broad style preferences. | Taste changes with occasion, season and budget. A fixed answer set can't be corrected in the moment. |
| Order history & re-buy prompts | Remembers purchases and nudges repeat or similar buys. | History is used to sell the next item, not to understand what the wardrobe already covers or is missing. |
| Visual search & similar items | Finds catalogue items that look like a photo or product you tapped. | Matches a single item in isolation, with no memory of it afterwards and no sense of the full outfit. |
Phase 1
Digital Wardrobe
Photograph what you own, or pull it in automatically from past Myntra orders. AI garment recognition tags category, colour and pattern so building the closet takes minutes, not an afternoon. Recommendations then mix owned pieces with catalogue items, so a suggested outfit can be three things you already have and one you don't.
Phase 2
Conversational Stylist
The static style quiz becomes a chat. Ask for an outfit for a specific occasion, say what you can spend, and get a look you can refine in follow-up messages. The stylist stays budget aware and uses shopping history, including returns, so it stops suggesting the fits and sizes that never worked.
"As Priya, a 27 year old marketing associate with a full wardrobe and no time, I want the app to know what I already own, so that I can be told what to wear on a Monday morning instead of being sold another kurta I don't need."
Feature scope
| Phase | Priority | Feature | Description |
|---|---|---|---|
| Phase 1 | P0 | Wardrobe upload | Add owned clothes by photo, plus auto-import of past Myntra orders. |
| Phase 1 | P0 | AI garment recognition | Detect category, colour and pattern so each upload is tagged without manual entry. |
| Phase 1 | P1 | Mixed outfit recommendations | Build looks that combine owned pieces with a small number of catalogue suggestions. |
| Phase 1 | P2 | Wardrobe gap view | Show what the closet is missing before recommending anything to buy. |
| Phase 2 | P0 | Conversational stylist | Chat replaces the static quiz: ask for an occasion, get a look, refine it in follow-ups. |
| Phase 2 | P1 | Budget aware suggestions | Respect a stated spend limit and lean on owned items when the budget is tight. |
| Phase 2 | P2 | History informed styling | Use past purchases and returns to avoid repeating fits and sizes that did not work. |
Rollout
Phase 1
Wardrobe upload and garment recognition to a small beta cohort, measured on items added per user and week-two return rate.
Phase 2
Conversational stylist replaces the style quiz for that cohort, measured on chats started and outfits saved.
Phase 3
Full rollout with budget aware and history informed styling, measured on styling driven add-to-cart and return rate.
Questions I would want to answer before building this
- Will people put in the effort to photograph clothes they already own, and what is the smallest number of items that makes recommendations feel useful?
- Can garment recognition be accurate enough on ordinary phone photos that tagging never feels like data entry?
- Does styling the existing wardrobe grow purchases or quietly replace them, and which metric decides that honestly?
Want the whole thing?
The PRD has the full feature landscape, personas, scope and metrics.
Read the Full PRD ↗