Smart Digital Platform for Personalizing Outfit
COMPANY
OVERVIEW
This project focuses on designing a smart, intuitive platform that enables users to discover, curate, and plan personalised outfits while organising and optimising their digital wardrobe. The goal is to create meaningful, recommendation-led interactions that help users build a stronger emotional connection with their clothing and personal style.
By reducing decision fatigue and making daily outfit selection effortless, the platform streamlines styling and simplifies shopping through partner retailers. Ultimately, it transforms the process of getting dressed into a more enjoyable, confident, and stress-free experience.
PROJECT
SCOPE
Through broad ideation and concept exploration, we designed a fashion-first wardrobe and shopping app that moves beyond generic recommendations, focusing instead on immediate value, personalized styling, and seamless wardrobe integration.
THE
GOAL
-
Design a minimal, fashion-forward platform that:
-
Delivers immediate value on first use
-
Helps users organise wardrobe and wishlist
-
Provides relevant, trustworthy recommendations
-
Supports real-life outfit planning
-
Encourages long-term engagement
THE PROBLEM
Users struggle to translate their wardrobe into complete, cohesive outfits, leading to decision fatigue and reduced confidence in both daily styling and purchasing decisions. They expect fast, personalised recommendations that reflect their body type, preferences, and individual sense of style.
However, existing solutions introduce friction through lengthy onboarding flows, unclear communication around data usage, and generic recommendations that lack relevance. In addition, users lack effective tools to plan outfits for different occasions or gain visibility into wardrobe insights such as item usage and cost-per-wear.
As a result, outfit selection becomes time-consuming and inefficient, wardrobes remain underutilised, and users feel disconnected from their personal style limiting their ability to make confident, informed decisions.
OPPORTUNITY
The opportunity is to reimagine the wardrobe as a smart, personalised styling assistant a “friend in your pocket” that helps users confidently put outfits together using what they already own.
By improving visibility, personalisation, and guidance, the app can reduce decision fatigue, support outfit planning for different occasions, provide insights like item usage and cost-per-wear, and seamlessly connect styling with relevant shopping options.
HOW DO PEOPLE
DECIDE THEIR
PERSONAL STYLE
RESEARCH
APPROACH
To develop a clear understanding of the problem space, we used a combination of qualitative and quantitative research methods. This approach helped uncover user behaviours, identify key pain points, and validate insights that informed the product’s design direction.
RESEARCH
GOAL
The research aimed to:
-
Understand how users currently discover and manage fashion items
-
Identify pain points in existing fashion and shopping apps
-
Explore how users make decisions about their personal style and purchases
-
Identify opportunities for improving convenience and decision-making
Method #1
Heuristic Evaluation
We evaluated existing products using usability principles to identify systemic issues.
Key Findings
-
Navigation complexity increased cognitive load
-
Poor feedback reduced user confidence
-
Feature-heavy interfaces created overwhelm
Insight: Simplicity and clarity would be critical differentiators.
Method # 2
User Interview + survey
Participants
We conducted 15 semi-structured interviews with participants aged 18–60, with the majority being younger users who frequently engage with fashion, shopping, and lifestyle apps.
Method
The interviews explored users’ current behaviours when discovering fashion items, managing their wardrobe, and deciding what to purchase. Participants were asked which apps they currently use, the challenges they experience with those tools, and what features would make a fashion app more convenient and useful for them.
Survey Validation
Insights from the interviews informed the design of a follow-up survey aimed at validating patterns and quantifying user preferences across a broader audience. The survey helped confirm recurring pain points and measure interest in potential features.
Outcome
This mixed-method research approach helped identify key user needs, validate design assumptions, and prioritise features that could address decision fatigue and improve the overall fashion discovery and wardrobe management experience.
COMPETITOR ANALYSIS
We analyzed four key competitors:
ALTA
DRESSLY
FITS
STYLE DNA
KEY
STRENGHTS
-
Strong personalisation systems are becoming standard
Most leading apps (e.g., Alta, Dressly, Style DNA) leverage user profiling, body analysis, and preference learning to improve outfit relevance. This increases perceived accuracy and user trust in AI styling outputs.
-
Utility-driven features that support real decision-making
Successful products extend beyond styling into daily context—weather, calendar integration, outfit history, and planning tools. This shifts the experience from “recommendation” to “decision support”.
-
Social discovery reinforces engagement loops
Platforms like Fits demonstrate that community-driven feeds, saved looks, and outfit sharing significantly increase retention by combining inspiration with validation loops.
-
Reduced-friction interaction patterns
AI chat, auto-saved outfits, voice input, and simplified onboarding reduce cognitive load and make styling accessible even for non-fashion users.
-
Structured organisation improves scalability
Tag-based systems (occasion, season, brand, category) enable users to manage large wardrobes and improve retrieval efficiency over time.
KEY
WEAKNESSES
-
Weak information architecture and navigation clarity
Many apps rely heavily on icon-based navigation and visually dense layouts, leading to ambiguity and high cognitive load for first-time users.
-
Fragmented and inconsistent user flows
Common issues include repetitive screens, non-intuitive interactions, and incomplete task flows that interrupt continuity in styling or browsing journeys.
-
Limited inclusivity in AI representation
Body and face modelling systems often fail to represent diverse body types accurately, reducing emotional resonance and perceived fairness of recommendations.
-
High dependency on manual input
Users are still required to upload, photograph, or manually tag clothing items, creating a significant barrier in wardrobe setup and ongoing maintenance.
-
Shallow AI interaction models
Most AI stylists operate as single-response systems rather than adaptive conversations, limiting contextual reasoning and long-term user understanding.
MARKET
GAPS
-
Shift from “AI recommender” to “conversational stylist”
Opportunity to design an adaptive AI that asks clarifying questions, compares options, and builds contextual outfit reasoning rather than outputting isolated suggestions.
-
Progressive onboarding instead of upfront commitment
Allow users to experience value before full setup, then gradually enrich personalisation through behavioural learning and optional input.
-
Inclusivity as a core system, not a feature
Differentiate through stronger body diversity modelling, improved avatar systems, and recommendation logic that adapts across different body realities—not just visual representation.
-
Low-effort wardrobe ingestion
Reduce friction through smart capture methods (e.g., image recognition, bulk upload flows, auto-tagging, or scanning-based input), shifting effort away from users.
-
Unified ecosystem across styling, planning, and inspiration
Most competitors split utility (planning), inspiration (social), and execution (outfits). A clear opportunity exists to merge these into a single continuous decision flow.
USER EXPECTATIONS
& WHAT I DESIGNED
“a fashion stylist app would be my trusted style companion if It has to be curated to me, and Everything is all about me.”
Design Decision
-
Implement a personalisation engine driven by onboarding inputs and user preferences
-
Surface user-specific content and recommendations across all key touchpoints
-
Continuously refine outputs based on user behaviour and interactions
Impact
-
Positioned the product as a trusted “personal stylist,” increasing perceived relevance and long-term engagement
“I feel seen, confident, & understood when styling suggestions respect my personal style and understand the practical needs of my daily life”
Design Decision
-
Design context-aware recommendation logic aligned with lifestyle, routines, and user intent
-
Incorporate practical constraints (occasion, weather, habits) into suggestions
-
Reduce reliance on generic outputs in favour of tailored experiences
Impact
-
Strengthened emotional connection and increased user confidence in styling decisions
“When I browse AI suggestions filter results by price to see more affordable options”
Design Decision
-
Introduce dynamic price filtering and budget-aware recommendation logic
-
Enable multi-criteria filtering (price, style, category) within search and discovery
-
Surface affordable alternatives without compromising relevance
Impact
-
Improved accessibility and supported faster, more confident purchase decisions.
“I would use scheduling event in calendar every single day, also if I finding out which clothes I haven't worn and I can't match and then maybe throwing them away.”
Design Decision
-
Develop an outfit planning system with calendar integration
-
Introduce wardrobe usage tracking to highlight underutilised items
-
Surface context-aware outfit suggestions based on upcoming events
Impact
-
Supported proactive planning and increased utilisation of existing wardrobe items
“Interactions about tracking quality, size and saving money make styling app encouraging for me. Having a feature where users can track how much they spend on each purchase, I Like 100% I love that idea.”
Design Decision
-
Introduce cost-per-wear and spend tracking analytics
-
Capture brand-specific sizing and quality data
-
Provide insight-driven feedback loops to inform smarter purchasing decisions
Impact
-
Enabled more intentional shopping behaviour and increased user trust in the platform’s value
“I find personalized styling advice extremely helpful especially when it's based on my existing habits, like suggesting "You wear high-waisted pants often. "
Design Decision
-
Leverage behavioural patterns to generate habit-based styling suggestions
-
Deliver actionable micro-recommendations (e.g. styling tips based on past behaviour)
-
Continuously adapt suggestions based on user interaction history
Impact
-
Increased engagement through relevant, actionable guidance rather than passive inspiration
USER EXPECTATIONS & WHAT WE DESIGNED
“I feel frustrated by the lack of cohesion in my wardrobe where nothing seems to go together, and I'm constantly challenged by the inconsistent sizing across different brands”
Design Decision
-
Introduce outfit composition tools to help users build cohesive looks
-
Provide brand-specific sizing intelligence to reduce inconsistency issues
-
Improve wardrobe organisation and compatibility visibility
Impact
Reduced friction in outfit creation and increased confidence in both styling and purchasing
“The pain point… apps push the cheapest or most popular items and they don't care about the material or the price or anything like that. They just push everything.”
Design Decision
-
Shift to a preference-driven recommendation model over popularity-based logic
-
Integrate quality, material, and relevance signals into recommendation ranking
-
Prioritise user intent over generic trends
Impact
-
Improved recommendation accuracy and reduced frustration with irrelevant content
HOW MIGHT WE...
HOW MIGHT WE ...
Effortless Outfit Discovery
How might we help users easily discover and assemble outfits tailored to specific occasions so they can feel confident and prepared?
Understanding Wardrobe
Valuable
How might we help users manage their wardrobe while tracking brand-specific sizing, cost-per-wear, wish-lists, and wear frequency so they can make smarter styling and shopping decisions?
Engaging
Styling Experience
How might we turn the styling journey into something fun, friendly, and encouraging?
Truly
Personalised Suggestion
How might we capture meaningful personal details and turn them into precise, tailored recommendations so they can avoid unwanted styles?
MVP
-
Translated key HMW insights into core product features focused on outfit planning, wardrobe intelligence, and personalisation
-
Introduced a simple outfit planning calendar to help users organise and plan looks for specific occasions in advance
-
Integrated wardrobe analytics based on usage patterns, enabling users to track item wear frequency and better understand wardrobe utilisation
-
Developed a dual colour system combining personal preferences with individual colour analysis (hair, skin tone, eye colour) to improve styling accuracy and confidence
-
Established the foundation for a smarter styling experience that helps users curate outfits more effectively, gain wardrobe insights, and receive better personal support from Dressi
BRAND IDENTITY DEVELOPMENT
FROM INSIGHT TO PRODUCT REQUIREMENT
When defining the Dressi brand, we first identified the core audience as younger users who value confidence, self-expression, creativity, and individuality. They want to feel seen, supported, and empowered in how they dress and present themselves. Beyond this segment, Dressi also serves users seeking convenience, confidence, and clarity in their everyday styling decisions, allowing the brand to remain flexible across a wider audience.
To reflect these insights, the visual direction was designed to balance warmth, empathy, and empowerment. Warm earthy tones establish a comforting and approachable foundation, while brighter accents, particularly pink, introduce energy and personality aligned with a younger, expressive audience. The colour palette was intentionally selected to remain inclusive and adaptable rather than tied to a single age group or style preference.
Imagery focused on authentic, confident, and diverse individuals to reflect the full spectrum of personal style, from minimal and understated to bold and expressive. This reinforces Dressi’s positioning as an inclusive platform that supports different fashion identities and encourages confidence in everyday style choices.
TYPEFACE
Typography was selected to strengthen both personality and usability.
Geller adds warmth, character, and expressive confidence to key messaging, while Inter provides clarity, readability, and simplicity for body content and everyday interactions.
This combination creates a balance between emotional connection and functional accessibility.
BRAND
IDENTITY
Across the main product experience, soft neutral gradients were introduced as foundational backgrounds to balance accent colours and maintain a warm, approachable feel. Pink accents and warm black were used consistently for hierarchy, emphasis, and contrast.
To further align the interface with the brand personality, sharp rectangular components were replaced with rounded shapes and softer edges. This design decision created a more inviting, human-centred experience and helped differentiate Dressi from conventional AI-driven platforms.
This integrated system of colour, typography, imagery, and shape language positioned Dressi as more than a styling tool, establishing it as a supportive and approachable digital companion within users’ daily routines.
_edited_edited_edited_edited_edited.png)
_edited.jpg)
ONBOARDING
Building Personal Style
A lightweight onboarding system designed to build personal style intelligence without blocking product access or increasing drop-off risk.
Emotion-led entry experience
-
Opens with warm, energetic visuals and a soft colour direction aligned with fashion and self-expression.
-
Establishes an immediate emotional connection before any data input begins.
Design decision:
Prioritised emotional anchoring over instructional onboarding to reduce cognitive resistance in early engagement.
Progressive, skippable onboarding flow
-
Users can skip any step and still access the core app experience.
-
Style questions are structured progressively to understand preferences, lifestyle, and fashion behaviour.
-
Optional body measurement input is introduced as an enhancement layer rather than a requirement.
Design decision:
Reduced forced commitment to improve activation rate and avoid premature drop-off.
Dual-layer personalisation model
-
Layer 1: subjective style preferences (taste, inspiration, lifestyle)
-
Layer 2: physical attributes (body measurements)
Design decision:
Separated behavioural and physical data capture to reduce overwhelm while still enabling long-term AI personalisation accuracy.
HOMEPAGE
Central Orchestration Hub
A central orchestration hub designed to minimise decision fatigue and enable fast, context-driven outfit planning.
Calendar-first structure for intent anchoring
-
Calendar highlights upcoming events and visually prioritises outfit needs.
-
Users can immediately start outfit creation if no plan exists.
Design decision:
Anchored the system around time-based context because outfit decisions are fundamentally event-driven, not browsing-driven.
Multi-entry outfit creation system
Users can start styling through:
-
Visual canvas creation
-
Wardrobe item selection
-
AI-assisted suggestions (Ask Dressi)
Design decision:
Designed for different cognitive styles (creative exploration vs structured planning vs assisted decision-making) instead of enforcing a single workflow.
Event-linked styling workflow
-
Users can directly connect outfits to specific upcoming events.
-
Enables quick editing and refinement based on schedule changes.
Design decision:
Reduced navigation friction between “planning” and “execution” by collapsing them into one loop.
Trend discovery with actionable pathways
-
Users can browse trends, share them, or open external product pages.
-
Items can be imported directly into the styling canvas for comparison with existing wardrobe pieces.
Design decision:
Converted passive inspiration browsing into actionable decision-making by connecting trends → try → compare → adopt.
WARDROBE
Mirroring Physical Wardrobe
A structured digital closet system that mirrors physical wardrobe behaviour while introducing intelligence-driven organisation and decision support.
Closet + Wishlist separation model
-
Closet represents owned items
-
Wishlist represents intent and future acquisition
Design decision:
Separated ownership vs aspiration states to improve clarity in wardrobe planning and reduce cognitive mixing of “have vs want”.
Styling-first closet environment
-
Users can mix and match outfits directly inside their wardrobe space.
-
Supports virtual try-on for validation before finalising looks.
Design decision:
Recreated physical dressing behaviour digitally to align with users’ existing mental model of “choosing from a closet”.
Smart item intelligence system
-
Each item includes metadata such as:
-
Usage frequency
-
Cost-per-wear insights
-
Size and fit tracking
Design decision:
Introduced behavioural feedback loops to shift decision-making from emotional selection to informed wardrobe optimisation.
Automated wardrobe population via receipts
-
Users upload digital receipts, and items are auto-detected and added to wardrobe.
Design decision:
Eliminated manual data entry to reduce friction and ensure wardrobe completeness over time.
WARDROBE DIGITAL RECEIPT
WARDROBE
VIRTUAL TRY-ON
WARDROBE
NEW OUTFIT
WARDROBE WISHLIST
ASK DRESSI
Conversational Ai Layer
A conversational AI layer designed to accelerate styling decisions and reduce uncertainty in outfit creation.
Intent-based styling assistant
-
Users can ask natural language questions such as:
-
Outfit suggestions for occasions
-
Styling compatibility for specific items
-
Provides contextual outfit recommendations similar to a personal stylist.
Design decision:
Shifted interaction model from navigation-based browsing to intent-driven conversation to reduce decision friction.
Guided onboarding prompts for first-time users
-
Predefined prompt suggestions help users understand how to interact with the AI.
Design decision:
Solved blank-state uncertainty by scaffolding initial interaction patterns.
DRESSI PROFILE
Centralized Personal Intelligence Hub
A centralised personal intelligence hub for managing identity data, advanced personalisation features, and system configuration.
Deferred setup management system
-
Stores onboarding data that users skip and allows completion later.
Design decision:
Decoupled onboarding completion from activation to avoid blocking early product value exposure.
Advanced personalisation engine inputs includes:
-
Body analysis
-
Colour palette generation
Design decision:
Introduced deeper biometric and aesthetic profiling to improve recommendation precision and differentiate from standard fashion apps.
Virtual avatar system for try-on
-
Users create a digital avatar based on collected personal data.
-
Used across try-on and styling simulation flows.
Design decision:
Strengthened continuity between user identity and outfit simulation to increase realism and decision confidence.
USABILITY
TESTING
ISSUES &
ITERATION
1. Users wanted more control before committing to onboarding
Insight:
Users preferred exploring app value before completing full personal details or setup.
Evidence: “I usually want to jump straight into what the app offers… before I bother filling in all the details.”
Issue: Mandatory onboarding created unnecessary friction for first-time users who wanted to validate product value early.
Iteration:
-
Added Skip Onboarding option for faster entry
-
Reduced required fields during initial setup
-
Allowed users to return later to complete preferences for more personalised recommendations
2. Homepage interactions created unmet expectations
Insight: Users expected interactive behaviour between trends, dates, and calendar content.
Evidence: “I expected that if I clicked the date, it would show my upcoming events & outfits on the calendar below.”
Issue: UI suggested clickable or connected elements without fully matching user expectations.
Iteration:
-
Improved affordance for clickable trend dates
-
Connected date selection to calendar filtering behaviour
-
Added clearer interaction feedback between trends and scheduled outfits
3. AI recommendations lacked conversational depth
Insight:
Users were satisfied with recommendations but expected more flexibility from AI.
Evidence: “The only thing is I would’ve liked more than one option, or even a follow-up question.”
Issue: AI delivered relevant outputs but interaction felt too linear and limited.
Iteration:
-
Expanded AI output to show multiple outfit recommendations
-
Added follow-up prompts for refinement (occasion, style mood, weather, colour preference)
-
Improved conversational flow to better mimic personal styling consultation
4. Core navigation validated successfully, with minimal friction
Insight:
Wardrobe, Wishlist, and Profile flows were consistently rated intuitive and easy to navigate.
Issue: No major usability blockers identified, but opportunity existed to preserve simplicity while scaling features.
Iteration:
-
Maintained existing navigation patterns
-
Prioritised consistency across icons, labels, and CTA placement as new features were introduced