Role
UX Researcher, UI/UX Designer
Deliverables
Wireframe, High Fidelity Prototype, Design System
Discipline
UX design, UI design, User Research, Usability Testing
Platform
Mobile
Timeline
10 Days
Tools
Figma
Overview
Frankly, I’m always forgetting about clothes that I purchased a while back and then proceeding to never wear it. It was a cycle I needed to break. Hence, I stumbled upon outfit planner apps online but found that many of the apps were either too big for my potato phone to handle or had features I knew I didn’t need. Moreover, none of them advocated for outfit repetitions, which was something I wanted to achieve.
In this project, I created a virtual closet app, Pakai, with the intention to highlight the basic features needed while encouraging users to repeat outfits and reuse clothes they never use.
Problem
I have clothes, but I don’t know what to wear
Users feel overwhelmed by their wardrobes and often forget or overlook items they already own.
Too much mental effort to plan outfits
Users struggle to plan their outfits ahead of time and find it hard to visualise outfits.
Existing closet apps feel overwhelming
Competitor apps often overwhelm users with too many features.
“How might we help users plan and wear more of their wardrobe effortlessly while avoiding outfit repetition?”
Solution
Easy sign up and login
A quick and intuitive login and sign-up flow, followed by a short intro to key features, helps new users get started with ease.
Upload clothes to your online wardrobe
Users can digitise their closet by uploading photos, with each item auto-categorized by Pakai AI and taggable by type, colour, or season for easy browsing and outfit building.
Create outfits from your closet
Once items are uploaded, users can visually combine clothing to create full outfits, helping users see how their clothes work together.
Curate themed lookbooks for any occasion
Users can group outfits into curated collections (e.g., "Work", "Vacation", "Date Night") using Lookbooks. These act as moodboards for inspiration and easy access.
Plan your outfits for the week
A visual calendar allows users to assign outfits to specific days. This supports intentional outfit planning and promotes usage of neglected items.
Smart outfit suggestions powered by Pakai AI
Using wear frequency and clothing tags, Pakai’s AI suggests old outfits and even new combinations from your closet. These looks encourage rotation and help revive underused items.
Process
Research
Before diving into design, I wanted to make sure I was creating a solution that truly met user needs, addressed common wardrobe struggles, and offered a meaningful improvement over existing virtual closet apps.
Hence, I focused on uncovering the following:
What are virtual closet apps doing well and what’s missing?
What struggles do users face with outfit planning and wardrobe management?
What are virtual closet apps doing well and what’s missing?
Key findings
AI integration is a common expectation but not always reliable
Most of the apps use some form of AI for features like outfit suggestion or automatic logging. As useful as they may be, these tools often underperform or feel inconsistent such as inaccurate tagging and AI generally not working as intended.
Many features yet often overwhelming
Most of these apps are packed with many features including outfit planners, a marketplace and social sharing. Yet, these extra features can feel extremely cluttered and overwhelming.
Clean and intuitive UI sets the standard
A clean and easy-to-navigate interface came out as strengths for these apps.
What struggles do users face with outfit planning and wardrobe management?
To further understand real user needs and pain points, I conducted a small survey via Google forms. With 5 respondents, I was able to gather a better landscape of what target users want from the app and frustrations about their wardrobe.
Closet disorganisation is a common pain point
Most users either rely on a mental catalogue of the clothes they own or don’t organise their clothes in any way. This then often leads to underused clothing items and frustrations on what to wear.
The need for assistance in outfit planning and smart organisation
All respondents expressed an interest in features like automatic clothing categorisation, smart outfit suggestions, reminders for under worn pieces, and weather-based planning.
Trust in AI is conditional and needs clarity
Most respondents are open to AI suggestions, but there are some hesitations around privacy, transparency and accuracy. Hence, clear communication on the workings of the AI and how it handles data is crucial for building trust.
Ideate
Brainstorming
Once I had a good amount of research on the background of this project under my belt, I spent some time brainstorming solutions and must-have features.
Crazy 8's
I wanted to specifically brainstorm ideas for outfit suggestions to encourage users to wear outfits they do not use often or have not thought of. Thus, I employed the “Crazy 8s” technique and came up with 8 different ways outfits could be suggested to users.
User Flow
According to the competitor analysis and key features decided during the brainstorming session, I also mapped out a tentative user flow.
Deciding
Following the idea generation phase, I analysed each solution and ultimately chose which of the ideas were more viable to execute in terms of design (and hypothetically in the development phase). I found that the first and second ideas (from top left) was the best choice in terms of implementation while staying true to the goal of simplicity.
Storyboard exploration
I then began exploring into structuring the app overall. Hence, I sketched out low-fidelity wireframes to get a better visual of the app’s layout. I used the low-fi wireframes to put together a storyboard so that I could get a better picture of how the screens would potentially flow.
Prototyping
Very quickly, I proceeded to work on the high-fidelity prototype. But before that, I created a design system to ensure the style of the app’s interface is consistent throughout my prototyping.
Design System
Hi-fi Prototype
The following picture showcases the created prototypes that only includes the initial screens that were involved in important processes to be tested with users later on.
I would also like to mention that I initially wanted to focus on a few screens that showcased the basic and crucial features but along the way, I felt that ignoring certain processes and screens could potentially hurt the user testing phase. Hence, it took me a little longer to come up with a complete prototype of Pakai.
Testing
Once the prototype was ready, I conducted remote usability testing with 2 participants, one with no familiarity with virtual closet apps and another that has used a similar app before. For easy referencing, I named them Newbie User and Experience User. These participants completed several tasks and provided feedback on the app’s functionality and features via a Google Form with scale and open-ended questions.
Key findings
The usage of AI was still vague
“Discover Looks” doesn’t make it obvious that these are AI-powered. I thought they were just curated collections.
—Newbie and Experienced user
Planner lacks clarity and control
There is no indication of how many outfits planned in a day at a glance. There is also a lack of control options such as easily removing and moving outfits.
—Newbie and Experienced user
Lookbook function needs clarity
"What’s the difference between a lookbook and the planner? One is by theme, one is by date?"
—Newbie user
Iterations
By addressing each of the problems discovered during the testing phase and as stated above), I refined the design where I could within the short timeframe.
Problem #1: The usage of AI was still vague
Problem #2: Planner lacks clarity and control
Problem #3: Lookbook function needs clarity
Learnings
Clear research questions matter
One respondent highlighted that a question about AI trust felt too vague, making it hard to answer honestly. This emphasised the importance of crafting research questions that are specific, contextual, and easy to understand to ensure that the insights gathered are accurate and actionable.
Not everything goes as planner
From revisiting design decisions to realising my initial timeline was a bit too ambitious, this project came with its fair share of challenges. I started out picturing a perfect execution, but I quickly learned that design is rarely linear. It’s a process of exploration, iteration, and learning from missteps along the way.
Designing with awareness and openness
While making assumptions in the initial start of the process is alright, I learned how important it is to not let my own biasness take control of the design. Keeping an open mind and validating ideas with real users helped ensure the solution (and myself) stay grounded.
Future
While working on this app, I had tons of ideas that I would have loved to implement but did not have the time to do so. Hence, here are my two cents on what I could have done to further the final design of Pakai.
Better onboarding
According to my competitive analysis and the user research I carried out, I noticed alot of users felt overwhelmed by existing similar apps or have not even used a virtual closet app before. Hence, I would have loved to implement a mini tutorial when a user signs up for the first time to ensure a smoother introduction to app.
More AI-powered functions
While the current AI feature focuses on outfit suggestions based on least-worn items, there’s room to explore smarter personalisation. In future iterations, I would love to add functions such as:
Outfit recommendations based on mood or calendar events.
Smart capsule wardrobe building tailored to a user’s lifestyle.
