Kleo
Designing a calmer, more actionable way for students and people who struggle with maintaining their spaces to organize, clean and care for their rooms.
Room health
A photo of the room becomes a readiness score and a set of specific improvements.
Room setup
Detected items, Kleo's additions and health history, all editable.
Today
An abstract goal reduced to the tasks that can be completed now.
Maintenance
Long-cycle upkeep surfaced before it turns into a problem.
Project snapshot
Maintaining a room can become overwhelming when the task feels too large or when people do not know where to start.
Students, hostellers and people who need more structured and actionable ways to manage their spaces.
Turn an overwhelming cleaning/organization task into small, understandable actions.
Kleo uses AI to understand the state of a room and convert it into personalized, actionable routines.
In short
Designed Kleo end-to-end — research synthesis, personas, information architecture, user flows and a high-fidelity AI room-organisation prototype.
Cleaning became one clear next action instead of an open-ended decision: capture the room, get a sequenced routine, do one thing.
Whether the single-action model lowers avoidance for people with ADHD across a two-week diary study.
This is an independent case study I designed to practice end-to-end product thinking — from research through to a working prototype.
View interactive prototypeProcess
I used a lightweight human-centered design process to move from understanding the problem and people to defining the opportunity, exploring possible solutions and refining the final product experience.
Understand the people, behaviours, environment and emotional challenges surrounding room organization.
Identify recurring pain points, motivations, barriers and opportunities.
Translate insights into product opportunities, flows and possible interactions.
Develop the information architecture, interaction model, visual language and interface.
Iterate on the experience and create a clearer, more approachable final product.
Understanding the problem
The problem is not simply that people do not clean. The experience of maintaining a space is shaped by how large the task feels, how clearly it can be started, and how much structure exists around it.
17.3%
Reported 30-day retention for productivity apps — a secondary-research data point that suggests tools in this category struggle to stay part of a routine. It supported the hypothesis that the barrier is sustained use, not intent.
Secondary research · not primary data
Factors that shape the experience:
Research
At first glance, the problem seemed simple: people struggle to keep their rooms clean and organized. Looking deeper, the challenge was less about knowing how to clean and more about how people perceive, start and maintain the task.
A messy room can become a combination of physical clutter, mental overload and emotional friction. When everything feels like it needs attention at once, even a simple task can feel difficult to begin.
How might we
How might we make taking care of a space easier to start — and easier to maintain?
The problem is cumulative
Rooms rarely become disorganized because of one large event. Small unfinished actions accumulate over time.
Small action
Clothes left on a chair
Another action
Objects placed wherever there is space
Another action
Trash postponed
Accumulation
Surfaces become cluttered
Perception
The entire room feels difficult to manage
Research insight
The experience is cumulative.
The product should support users before the room reaches a point where the entire space feels overwhelming.
Design implication
Support small, repeatable actions instead of relying only on large cleaning sessions.
The starting problem
Where do I start?
01
Too many possible actions
02
Decision overload
03
Delayed action
Insight
The barrier is often initiation, not capability.
People may know how to clean their room but still struggle to decide what to do first.
Design implication
Reduce the number of decisions required before action.
Psychological friction
One large block
“Clean your room”
“Clear your desk”
“Put away 5 items”
“Throw away empty bottles”
Behavioural loop
Mess creates overwhelm, overwhelm creates avoidance, and avoidance creates more mess. The loop keeps feeding itself, which is why the entry point matters more than the effort.
Self-reinforcing loop
MessOverwhelmAvoidanceMore messMore overwhelmAvoidanceInsight
A task can become harder simply because it is perceived as too large.
Design response
Break broad goals into small, concrete actions that feel achievable individually.
The emotional layer
Physical
Cognitive
Emotional
The state of a room can influence how a person feels about the space — and those feelings can make the task harder to begin.
Insight
The product needs to support the user’s emotional state, not make them feel judged for it.
Design implication
Kleo should feel supportive and non-judgmental.
Conceptual UX writing example
Avoid
“Your room is a mess.”
Use instead
“Let’s start with one small thing.”
Motivation vs maintenance
One-time cleaning
Maintenance
Insight
The product needs to support maintenance, not just recovery.
A successful cleaning session does not necessarily solve the underlying problem if users cannot maintain the new state.
Design implication
Move from a one-time cleaning experience toward lightweight recurring routines.
Progress
Cleaning can feel unrewarding because the effort involved is often temporary and difficult to quantify. Making progress visible can create a stronger sense of completion.
Before
Messy / many tasks
Action
Complete a small task
After
One area improved
Progress
More tasks completed
Confidence
The room feels more manageable
Conceptual room-health indicator. It is a product feature for communicating state and momentum — not a measured or validated research metric.
Structure without control
Too much structure
The app becomes another task
Lightweight structure
The app supports the task
Insight
The product should provide structure without requiring users to manage the structure itself.
Design implication
Prioritize one clear next action over a long list of everything that needs to be done.
Room care is multidimensional
Different room problems require different actions. Treating everything as one generic cleaning task makes the experience harder to understand.
Cleaning
Organization
Maintenance
Enhancements
Room health
This became the foundation for the product structure — four dimensions rolling up into one readable state of the room.
Simplified Kleo dashboard categories
Where AI actually helps
Observe
Understand the room
Understand
Identify areas requiring attention
Prioritize
Determine what matters first
Recommend
Suggest an actionable next step
Without AI
With Kleo
AI is useful when it turns understanding into action.
The point is not that AI replaces the user — it removes unnecessary cognitive work before the user has to do anything.
Research synthesis
People avoid large cleaning tasks.
→The task feels overwhelming.
→Break cleaning into small actionable steps.
People struggle to decide where to begin.
→Too many possible actions create decision fatigue.
→Give users a clear recommended next action.
People complete cleaning but struggle to maintain it.
→The room gradually returns to its previous state.
→Create lightweight recurring routines.
People can feel frustrated or guilty about their environment.
→Negative emotions can reinforce avoidance.
→Create a supportive and non-judgmental experience.
Users cannot always see the value of small actions.
→Progress feels invisible.
→Make progress visible and tangible.
Users need to understand what is wrong with their space.
→Analysis itself can require effort.
→Use AI to turn observation into prioritized actions.
The design opportunity
Make room care feel manageable.
Rather than building another cleaning checklist, Kleo could act as a personal room-care companion that helps users understand their space, decide what matters now and take one manageable action at a time.
Make it actionable
Turn overwhelming goals into clear next steps.
Make progress visible
Help users see that small actions are creating change.
Make maintenance easier
Support sustainable routines instead of occasional deep-cleaning sessions.
Research into product
“I don’t know where to start.”
→Recommended next action
“Everything feels like a lot.”
→Small actionable tasks
“I clean it, but it gets messy again.”
→Lightweight routines
“I want to know if I’m actually making progress.”
→Room health + progress
“I don’t know what needs attention.”
→AI room analysis
These insights shifted the problem from “help people clean” to “help people know what to do next.”
Next · Information architecture → user flow → design
Personas
Two personas held the research in place while the product took shape — one built around structure, one around motivation.
Lives in a hostel and balances classes, assignments and social life. She wants a clean room but postpones cleaning because it feels overwhelming, and she rarely knows where to start.
Has ADHD and experiences decision paralysis around large cleaning tasks. He needs work broken into small steps and finds most productivity apps too complicated and demanding.
Affinity mapping
Statements collected from community discussion and secondary research were clustered until the same five problems kept reappearing: starting is harder than cleaning, large tasks cause avoidance, routines collapse, existing apps add work, and progress is what keeps people going.
User flow
The core flow reduces the number of decisions users need to make before taking action. Instead of asking users to plan an entire cleaning session, Kleo moves them from capture to analysis to a routine, and then to one manageable next step.
Welcome and onboarding, room capture, AI recognition and correction, room health and issue categories, routine creation, and the dashboard loop with progress verification.
Design direction
Because the research pointed at overwhelm and decision fatigue rather than laziness, the interface had to lower effort before it did anything else. It intentionally avoids the density of a generic productivity dashboard.
Weight and spacing carry the structure instead of hard containers.
A friendly register that never reads as a productivity dashboard.
Each action is a self-contained, scannable unit.
Improvement in the room is reflected back to the user.
One primary action per screen, sized to be finishable.
Fewer decisions before the first useful step.
Design system
Josefin Sans carries the whole type scale, a three-colour system separates primary actions, warnings and confirmations, and soft badge tints code priority without shouting. The mascot keeps the tone non-judgemental across every screen.
Wireframes
Low-fidelity wireframes mapped out the core screens and information hierarchy, keeping the focus on clarity and task flow before introducing colour, type and final UI details.
Design decisions
Five decisions carry most of the concept. Each one traces back to a pattern found in the research.
The room health score gives users an immediate understanding of their current state and creates a visible sense of progress.
Breaking the experience into cleaning, organization, maintenance and enhancements makes a large problem easier to understand.
A clear primary action reduces decision-making and gives users an obvious entry point.
The Today view converts an abstract goal into concrete actions that can be completed now.
AI helps users understand their room without requiring them to manually assess everything themselves.
Final product
The product reads as one sequence: onboard, photograph the room, correct what the AI found, turn it into a routine, and then work through it a task at a time.
Onboarding asks only what changes the recommendations: whole home or one room, starting fresh or already set up, when free time usually happens, how often the user wants to tidy and how much time is realistic. Five light questions replace a setup wizard, and the mascot carries the tone so the first minute feels like a conversation rather than a form.
Instead of describing the mess, the user photographs it. Four photo slots, an optional voice note and a typed prompt feed the same analysis. Kleo then reads the images back — the items it recognised, grouped by category — and asks the user to confirm, correct or add anything it missed, so the AI stays a draft the user approves rather than a verdict.
The analysis becomes an editable room. A room health score gives an immediate read on the current state, and the findings split into cleaning, organization, maintenance and enhancement issues with priority tags. Every detected item can be renamed, resized, removed or added, which keeps the user in control of the data the routine is built from.
Each category becomes a short, reviewable list rather than one large chore. Items carry a priority tag and a suggested frequency, and a single primary action — save and add to my routine — turns a set of observations into a schedule. Nothing is committed until the user accepts it.
The dashboard opens on Today: the current date strip, the tasks due now grouped by priority, and a bottom navigation of Today, Rooms, Calendar and Profile. Completing a task can be closed with a progress photo, and Kleo compares it to the earlier image to report what improved and what still needs attention — so effort is tied back to the state of the room instead of a streak counter.
Before / after
The earlier direction asked the user to plan before the product gave anything back. The refined direction inverts that.
Before
After
What changed
These are qualitative design improvements from a self-directed concept. The next step is putting them in front of real users to see how they hold up.
Reflection
The challenge was not simply creating a cleaning checklist; it was designing around motivation, overwhelm and maintenance.
Breaking down a large task can make the experience feel much more approachable.
AI should reduce effort and decision-making rather than become another layer of complexity.
The visual design becomes more effective when the information architecture and interaction model are clear first.
Research findings
Secondary research and recurring patterns in public community discussion about cleaning and organization.
Design decisions
Choices made in response to those patterns — structure, categories, and the single-action model.
Conceptual outcomes
Kleo is an independent case study I designed to practice end-to-end product thinking — research through to a working prototype. Real-world testing is the next chapter, not a missing one.