A robot that gives kids a way out
WonderBudi is an AI robot designed for elementary school classrooms. It gives children a safe, engaging space to calm down during emotional distress, while silently alerting teachers and parents when a child needs further professional support.
The system pairs a biometric wristband called the Budiband with a child-facing robot and a parent-facing mobile app. When a child’s vitals signal distress, the ecosystem activates: teachers are notified, the child is walked to WonderBudi, and a structured 20-minute session of play therapy activities begins. At the end of every session, a single emoticon check-in determines whether the child returns to class or meets with a school counselor.
My Role
UX Researcher · Visual Designer · Prototype Lead · User Testing Lead
Technologies
Biometric Sensing (O2, heart rate, skin temperature) · NLP Voice Interaction · OLED Touchscreen Interface · Mobile Parent Dashboard · Card Sorting and Tree Testing
Three components of the WonderBudi ecosystem working together: the robot, the Parent App, and the Budiband wristband
Children who cannot speak up
Mental health challenges in children are widespread and frequently go unaddressed. Young children lack the maturity and vocabulary to identify what they are feeling, which makes it nearly impossible for them to ask for the help they need.
Schools are structurally underprepared to close this gap. There are not enough counselors to monitor every student, and most children will not surface distress on their own. By the time a problem becomes visible, it has often gone unaddressed for far too long.
Help children calm down in times of distress and direct them to mental healthcare services when they do not have the ability to express their need for it on their own?
Secondary research: childhood trauma is common, but school mental health resources reach fewer than a third of students nationally
1 in every 5 children aged 5 to 17 currently needs mental health support
A counselor’s perspective: the bottleneck is identification, not treatment capacity
Four recurring problems identified across initial interviews and secondary research
Listening before designing
A broad research effort spanning secondary literature, two surveys, and 16 structured interviews across four participant groups produced 22 core insights that shaped every design decision that followed.
16 interviews across four participant groups yielded 22 actionable insights
Survey Findings
Three core themes from the child-focused survey: expression, boredom, and embarrassment
School counselors flag identification as the core challenge; social workers flag the need for play-based therapeutic approaches
Voices from the Field
School counselors and child psychologists both point to identification as the hardest part of delivering support to children
Children with trauma describe boredom with traditional therapy and an inability to articulate what they feel
50 out of 52 parents said they would allow their child to wear a biometric device at school
Four overarching insights consolidated from the full survey dataset
Secondary Research
Legal context: mandated reporting obligations and school compliance requirements shaped the Parent App’s notification and consent design
Secondary research confirmed the gap between need and access in public school mental health support
Affinity Mapping
763 data points from all 16 interviews organized by theme and participant type
A 6-step affinitization process from raw data to three How Might We statements
The overarching How Might We that unified all subsequent design work
The Pivot
Mid-research, the team discovered that schools did not need more therapy programs. The real problem was identifying which students needed those programs in the first place. This drove a strategic pivot that redefined the entire product direction.
The pivot: from helping children cope directly, to helping the adults around them identify who needs support before it is too late
From insight to system
With the pivot established, the team built personas, mapped emotional user journeys, ran a Crazy 8 ideation sprint, and selected the WonderBudi ecosystem as the direction most capable of serving children, parents, and counselors simultaneously.
Personas and User Journeys
Tommy (child) and Peter (parent): two perspectives on the same unmet need, separated by a communication gap WonderBudi is built to bridge
Tommy’s emotional arc across a school year, with WonderBudi as the turning point where the downward spiral is interrupted
Peter’s journey: a parent who cannot reach his child, until WonderBudi gives him a window into his son’s school day
Ideation
Crazy 8 ideation: eight rapid concept sketches generated in eight minutes to open up the solution space
Concept 1 (WonderBudi) was selected over Concept 2 (virtual counseling) because Concept 2 raised access, stigma, and ethics concerns
A deliberate scope boundary: WonderBudi is an early-intervention tool, not a replacement for ongoing therapy
Design constraints clarified before entering final concept development
Four expected outcomes that served as success criteria throughout the project
Storyboard
A classroom storyboard placed WonderBudi in a realistic school scenario, showing how three children with three different emotional states are all served by the same system on the same day in Mrs. Zimmerman’s class.
Three children, three situations, one classroom: a realistic scenario to stress-test the WonderBudi system
The Budiband detects what teachers cannot always see, and surfaces it before the moment is missed
All three children use the same WonderBudi robot, each choosing their own activity to decompress in their own way
A single emoticon check-in routes each child correctly: back to class, counselor visit, or continued monitoring
Building the WonderBudi world
The final design is a connected three-part ecosystem: the WonderBudi robot, the Budiband biometric wristband, and the Parent App. Each component serves a different user, but all three share a data layer that keeps counselors, teachers, and parents informed in real time.
The WonderBudi robot, Budiband wristbands in five colors, and the tap-to-start onboarding screen
The Ecosystem Flow
Before WonderBudi: a consent and onboarding process ensures parents are informed and in control before their child’s first session
From biometric alert to classroom return: the full in-school flow across nine structured steps
The WonderBudi Session
Steps 2 and 3: personalization through avatar selection followed by an opening emotional check-in
Steps 4 through 8: structured play transitions into a guided breathing moment and reflection before the child returns to class
The activity screen lets each child choose what calms them most, personalizing every session without requiring adult input
Voice Interaction Scripts
Three conversation scripts cover the most common emotional states children arrive with. Each follows the same structure: an opening check-in, a play session, a guided reflection, and a closing affirmation that the child can return anytime.
Script 1: a child feeling sad gets a reflection conversation that ends with a concrete social next step
Scripts for anger (left) and anxiety (right): each emotion gets a tailored opening, then converges on the same reflective structure
Hardware and Components
Robot components (left) and the Budiband wristband (right): no screen on the band ensures children are never distracted by their own biometric data
Budiband in detail: comfortable, waterproof, screenless, and available in five colors for personalization
Parent App
Parents have 24/7 access to their child’s biometrics and WonderBudi session history through the companion app. The dashboard acts as a single control center: showing vitals, session transcripts, counselor notes, and activity permissions together in one view.
The Parent App gives families a continuous window into their child’s wellbeing throughout the school day
Session history with counselor notes (left) and long-term biometric trends (right) give parents a complete and evolving picture of their child
The dashboard acts as a control center (left); settings give parents full authority over which activity types their child can access (right)
Business Model Canvas: a two-tier pricing model makes WonderBudi accessible to schools at different budget levels
Why WonderBudi works the way it does
Four decisions defined the character of the product. Each one came directly from something a real user said or did during research or testing.
Early concepts included a display on the wristband to show the child their own biometric readings. Research with child psychologists showed that giving children real-time access to their own stress data increases anxiety rather than reducing it. Removing the screen entirely kept the band invisible to the child while still giving parents and teachers full visibility through the app. The band works best when the child forgets it is there.
Play therapy literature and direct interviews with school counselors pointed to 15 to 20 minutes as the window where a child can decompress without missing enough class to fall behind. Shorter sessions did not give enough time for the reflection conversation. Longer sessions created re-entry anxiety. The 20-minute structure was not arbitrary. It was the clinical consensus from the professionals interviewed during research.
A 3-option check-in (happy, neutral, sad) was the first design. Counselor feedback in testing showed that a 3-point scale collapsed too many emotional states into the same category and routed children incorrectly. A 10-point scale with a color gradient from red to green gives the system enough resolution to distinguish a child who is slightly upset from one who needs immediate counselor intervention. The routing accuracy depends on the granularity of the input.
Placing WonderBudi in the classroom itself was the original plan. Teacher interviews revealed that a robot in the classroom would create distraction and stigma, with other students watching when a classmate was sent to use it. Moving it near the counselor’s office normalized the visit as a routine school activity rather than a visible disciplinary or emotional event, protecting the child’s dignity during a vulnerable moment.
Testing with real educators and children
Three rounds of user testing validated the Parent App’s information architecture through card sorting and tree testing with children and parents, and gathered direct feedback from six real educators and counselors across international and domestic schools.
Six educators and school counselors across international and domestic schools participated in structured testing sessions
Four complementary testing methods: card sorting and tree testing for architecture, interviews and observation for real-world behavior
Test 1 and 2: Card Sorting
Closed card sort (left) validated existing categories; open card sort (right) surfaced how parents naturally group app content
Two key findings: the WonderBudi label caused confusion, and users rely on concrete keywords rather than brand terms to navigate
Card sorting data pointed clearly to a 5-category navigation structure for the Parent App
Test 3: Tree Test
Tree test results: 90% task success rate and 75% directness confirmed the 5-category navigation structure was intuitive and ready for production
90% of participants reached the correct destination, confirming the 5-category structure mapped closely to how parents think about monitoring their child.
75% of answers were reached without backtracking, a strong signal that the information architecture was intuitive without prior training or guidance.
The term WonderBudi as a navigation label confused participants and was categorized separately. Labels were revised to reflect specific actions rather than the product name.
Users scanned for concrete keywords rather than brand names. Labels across the app were revised to prioritize action-oriented language throughout.
A system that speaks for those who cannot
WonderBudi demonstrates that a child who cannot yet put their distress into words does not have to stay invisible. The six educators and counselors who participated in testing responded with consistent enthusiasm. One 4th grade teacher described it as the first tool she had seen that could catch the kids who fall through the cracks before they disappear entirely. A school counselor at an international school noted that the emoticon routing system solved the identification problem that consumes the majority of a counselor’s day. By reading biometrics, offering play-based calm, and routing real data to the adults who need it, WonderBudi creates a bridge between a child in need and the professional who can help.
The final WonderBudi product: robot, Budiband wristbands in five colors, and the embedded activity touchscreen
A 90% tree test success rate and 50 out of 52 parents approving wearables for their children validated both the information architecture and the core premise: parents actively want visibility into their child’s school-day wellbeing and will adopt a system that gives it to them.
Seven core values guided every design decision, from the script language Budi uses to the color options on the Budiband
Four social change impacts: WonderBudi serves as an early intervention tool and a normalizer for children’s emotional expression at school
Manufacturing specification for production at scale, with two body material options depending on production volume
Future Directions
The longer WonderBudi knows a child, the more precisely it can tailor activities and conversation scripts to what works for that specific child over time.
A dedicated counselor view to see cross-class distress patterns, flag recurring students, and coordinate follow-up sessions without reviewing individual parent app logs.
A lower-cost tablet-only version at $300 would make WonderBudi accessible to schools with tighter budgets and no space for a dedicated robot station.
A multi-semester study tracking emotional regulation outcomes would provide clinical evidence to support adoption across broader school districts and funding bodies.