Loading

product/interactive designer

Project 03Applications

← All work

Lixil Toilet Cloud

Helping cleaning teams know what needs attention next.

必要な手入れを、必要なときに。The right care, at the right time.

Lixil Toilet Cloud is a mobile operations app for frontline restroom cleaning teams in Japan. It transforms usage data, equipment signals and cleaning records into clear, prioritised work.

The system already knew a lot about each restroom. The real UX problem was helping workers decide: what should I do next?

Role
UX/UI designer
Year
2022
Scope
Research · UX strategy · Information architecture · User flows · Interaction design · UI design · Design system
Context
Client work for LIXIL on Toilet Cloud, a smart-restroom service for the Japanese market - the real app's screens and flows, shown here in English
A hand holding a phone showing Lixil Toilet Cloud's Toilet usage screen - men's, women's and accessible restrooms, each booth a toilet icon ringed when in use, with how many are free - against a soft white wall

01The challenge

More data wasn't making frontline work simpler

Cleaning is physical work.

Workers move between floors, carry equipment, respond to interruptions and inspect real conditions. The phone is used briefly, often one-handed and sometimes while wearing gloves.

But the product was organised around system records rather than how cleaners actually work. A single visit to one restroom might involve three jobs.

  • Cleaning · Toilet 02
  • Refilling · Soap
  • Inspecting · Urinal 01

The system treated these as three separate tasks. The worker experienced them as one visit.

  1. 01

    Too much interpretation

    The system showed conditions, but workers still had to decide what deserved attention first.

  2. 02

    Fragmented work

    Multiple jobs in the same restroom required repeated interactions.

  3. 03

    Weak continuity

    Unresolved work and shift records were separated, making handover dependent on memory or manual notes.

02The key reframe

Design around the visit, not the database record

The most important change was redefining the unit of work. Instead of asking workers to navigate the system's data structure, Lixil Toilet Cloud groups related work around the place where that work actually happens.

System model
TaskTaskTask3 steps
Worker model
One visitSeveral tasksOne report3 steps

Core insight

Workers didn't need more information. They needed clearer decisions.

03UX strategy

Make the next action obvious

North star

Help the right work happen at the right place, at the right time.

必要な手入れを、必要なときに。

Principles

The redesigned experience follows four principles.

  1. 01

    Prioritise

    Surface what requires attention now instead of presenting every restroom equally.

  2. 02

    Batch by context

    Group work that can be completed during the same physical visit.

  3. 03

    Recommend, don't command

    Sensor data can suggest what might need attention, but the worker standing in the restroom still has the final judgment.

  4. 04

    Preserve continuity

    Unresolved work should naturally move into the next shift without requiring someone to rewrite it.

04Core experience

An action-first home

The previous mental model started with facility data. Lixil Toilet Cloud starts with work.

The Home screen separates the shift into three levels, so workers see priority before the underlying data.

  • Urgent

    • Possible blockage

      May require immediate attention

  • Next recommended

    • 5F Women's restroom

      Suggested by usage patterns

  • Later

    • Soap refill

      Routine, can safely wait

Design principle

Answer "What deserves my attention first?" before showing "What is happening in the facility?"

User flow - one shift
Start shiftHomeUrgent · Next · LaterAnythingurgent?IncidentPossible blockageIssueconfirmed?EscalateNote what you sawService visit3 tasks · one placeStart visitComplete each taskAll tasksdone?PostponeShiftover?Shift logDone · postponed · openHandoverYesYesNoNoNoYesNoYesLocation, fixture andtime come pre-filledLocation, worker andtime captured onceOpen items carrytheir history
  • Screen
  • Decision
  • Action
  • Yes
  • No
Three phones in front of a white-tiled restroom showing Lixil Toilet Cloud's log-in, home and side menu screens; home lists toilets to clean with Urgent, Next up and Overflow tags, a saved draft and Clean buttons
Fig. 01Home - the action-first screen

One visit, several tasks

When several jobs belong to the same restroom, Lixil Toilet Cloud combines them into a service visit. Location, worker and timestamp are captured once when the visit begins.

5F Men's restroom

3 tasks

  • Toilet 02 · Cleaning
  • Soap · Refill
  • Urinal 01 · Inspection
Start visit

Individual tasks still keep what's relevant to their type

  • Cleaning

    • Condition after cleaning

  • Refill

    • Resulting supply level

  • Inspection

    • Observed condition or issue

Design rule

Batch by context. Report by task type.

12:30

Home

Toilet 2

Urgent

Hikari > 5F Men's restroom
> Booth 2

03/2508:55

Toilet 4

Next up

Hikari > 4F Women's restroom
> Booth 4

03/2509:20

Toilet 1

Overflow

Hikari > 3F Men's restroom
> Booth 1

03/2509:42

Toilet 3

Sora > 2F Women's restroom
> Booth 3

03/2510:05

Toilet 6

Sora > 1F Men's restroom
> Booth 6

03/2510:31

Toilet 5

Hikari > 2F Women's restroom
> Booth 5

03/2511:10

Step 1 / 5

Clean

Open the report for the first toilet.

Or try Select to clean several at once.

Fig. 02Service visit - a working prototype. Report one toilet, save a draft, or select several and clean them together; finished work moves to Work history.

Exceptions become unmistakable

Routine work should feel quiet. An incident shouldn't. When Lixil Toilet Cloud detects something unusual, such as a possible blockage, the experience shifts from passive monitoring to guided action.

  1. Check the condition
  2. Confirm the issue
  3. Escalate to maintenance

Known information such as location, fixture, detection time and issue type automatically moves into the maintenance request. The worker contributes only what the sensor cannot know: what they actually saw.

Design principle

An exception should change the interface's volume, not its vocabulary.

08:55

Alert

Log

Toilet 2

Check needed

Hikari > 5F Men's > Booth 2

08:55 detected · Kenta Sato

Possible blockage or overflow

A sensor detected a problem. Please check on site.

Where to check

Booth 2, 5F men's restroom. Check the toilet's water level and flow.

Assigned to

Kenta Sato · On shift

Step 1 / 6

Check on site

A sensor flags a possible blockage. It starts a check - it doesn't confirm a fault.

Fig. 03Incident mode - a working prototype. The alert starts a check, not a repair; confirm it to send a pre-filled request, or record that nothing was wrong. Sent stays unresolved until maintenance fixes it.

Handover happens naturally

Reporting should be a consequence of doing the work - not another task at the end of the shift.

The Shift Log automatically captures completed, postponed and unresolved work. When the shift ends, open items move to the next worker with their context and history attached. No rewriting, no separate end-of-day report, no dependence on memory.

Design principle

If the system watched the work happen, it shouldn't ask the worker to describe it again.

08:00

Shift log

Let's start your shift

Check who's working and where.

Worker

Kenta Sato

Using this device

Location

Hikari

Your area · 3F–5F

2 items from the last shift

See postponed work and repairs in progress.

Step 1 / 5

Review handover and start

Check the worker and area, then review the handover

Tap the glowing button on the phone.

Fig. 04Shift log & handover - a working prototype. Tap the main button to run a shift; the last step hands the open work to the next worker.
16:00

Handover

Kenta Sato · Hikari

08:00 – 16:00 · 8 done

To hand over · 2

Soap refill

Postponed

Hikari > 3F Men's

Out of stock. Refill after the delivery.

15:30 · Kenta Sato

Toilet 2

In repair

Hikari > 5F Men's > Booth 2

Blocked. Check on the repair with maintenance.

In repair · M-204

Place, what was done and the next step fill in automatically.

Step 1 / 3

Choose recipient

Open work carries its history forward - choose who gets it.

Add a note puts a note on an item first - only if something is missing.

Fig. 05Outgoing handover - a working prototype. Open work carries its history forward; add a note only if something is missing, choose who receives it, and the shift ends once it is sent.

05UI & accessibility

Built for quick recognition

The interface is designed to be read for a few seconds at a time while standing and working.

  • Interaction

    • 52–56px primary controls
    • Large touch targets
    • One-handed interaction
    • No precision-dependent gestures

  • Status

    • Never relies on colour alone
    • Icon + label + colour communicate urgency and state

  • Language

    • Familiar workplace language, not software terms
    • “Finish cleaning”, not “Complete task”

Accessibility

Accessibility isn't a separate "senior mode". Readable typography, strong hierarchy, explicit status and forgiving touch targets are part of the default experience.

12:30

Toilet usage

Select
3F - NorthFree:Men's 7/10Women's 7/10Accessible 7/10
4F - North
5F - South

Step 1 / 5

Men's

Each floor shows how many toilets are free right now.

Tap a group to see every fixture.

Fig. 06Usage status - a working prototype. Rings read at a glance: dark means heavy use since the last clean. Tap one to clean it and watch it reset; Occupancy shows what's occupied, and for how long.

06What changed

From monitoring restroom data to helping people act on it

  1. 01

    Before: Individual fixture records

    Lixil Toilet Cloud: Prioritised operational work

  2. 02

    Before: Manual multi-select

    Lixil Toilet Cloud: Contextual service visits

  3. 03

    Before: One completion pattern

    Lixil Toilet Cloud: Task-specific reporting

  4. 04

    Before: Sensor alerts

    Lixil Toilet Cloud: Guided incident response

  5. 05

    Before: Separate work records

    Lixil Toilet Cloud: Shift log + handover

  6. 06

    Before: System instructions

    Lixil Toilet Cloud: Explainable recommendations

  7. 07

    Before: Data-first interface

    Lixil Toilet Cloud: Action-first interface

07Intended outcome

What I would validate

I don't have measured results to share, so I don't claim performance improvements. Instead, I would validate whether the new model creates:

  • Less unnecessary work

    Fewer visits to restrooms that don't require attention.

  • Faster response

    Clear separation between routine work and abnormal conditions.

  • Better continuity

    More unresolved work successfully transferred between shifts.

  • Higher trust in automation

    Workers understand why something is recommended and can override it when reality differs.

08Reflection

Two ideas I still design by

  • Design around real-world activity

    Several database records can represent one real human activity. Understanding that difference can remove entire layers of unnecessary interaction.

  • Automation should support expertise

    The strongest experience wasn't the one where software made every decision. It was the one where the system handled complexity quietly while leaving professional judgment with the person doing the work.

Takeaway

Good operational technology should disappear into the work.

現場に必要なのは、より多くの情報ではなく、次に何をすべきかがわかること。What the field needs isn't more information - it's knowing what to do next.

More work