← SELECTED WORK / 01
SOLBIT · EDUCATIONAL COMPANION ROBOT

A little circle.
A whole new
language.

Designing a voice-first learning companion through projected light, conversation, and movement.

FROM EARLY PROTOTYPES TO A CONNECTED LEARNING EXPERIENCEMeet Solbit HUMAN–AI INTERACTION / CHINA
OBJECT 001 / SOLBIT
White Solbit robot with a large round head and three wheeled supports
HELLO, I’M SOLBIT.
COME A LITTLE CLOSER.
TRY AN EXPRESSION

A quiet circle. Ready to listen.

DRAG TO ROTATE · EXPRESSION STUDIES

MY ROLEProduct experience designer · Sole in-house designer
MY SCOPEExperience, physical product collaboration & brand
ATNanjing Perception Robotics
PERIOD2018–2020
MY ROLE / RESPONSIBILITIES & COLLABORATION

One designer.
Across the
whole experience.

As the sole in-house designer, I connected Solbit’s product experience, physical design, brand, and communication.

I worked closely with a freelance industrial designer, the CTO, content writers, and translators to bring the product and its story together.

01 / EXPERIENCE & INTERACTION

How Solbit responds.

I designed the experience architecture across the robot, parent portal, and teacher workflows, including the projected expressions and interaction states that supported children’s conversations with Solbit.

02 / PHYSICAL PRODUCT & PROTOTYPING

From form to prototype.

I collaborated closely with a freelance industrial designer on product design decisions and CMF—colour, material, and finish—ideas. I also worked on 3D-printing the initial prototypes to explore the physical product.

03 / BRAND & MARKETING

A consistent identity.

I designed Solbit’s branding and marketing materials, extending the product’s identity across the touchpoints used to introduce it to different audiences.

04 / PITCHES & INVESTOR COMMUNICATION

Making the vision clear.

I designed pitch decks and investor materials, working closely with the CTO, content writers, and translators to shape the designs and communicate the product story.

01 / A DIFFERENT KIND OF COMPANION

Less tapping.
More talking.

We wanted children to interact with a companion through conversation, movement, and curiosity.

Solbit was an educational robot designed for children in China, including those growing up without siblings. It could recognise objects, respond to basic maths questions, move around the home, and return to its charging dock.

Its role was to support early learning alongside teachers and parents. The challenge was to make the robot’s responses understandable without a touchscreen.

NO TOUCHSCREENVOICE + PROJECTED LIGHTLEARNING THROUGH INTERACTION
Solbit in action.Watch on YouTube ↗
02 / THE DESIGN DECISION

One shape.
Many ways
to respond.

A glowing teal circle became Solbit’s visual vocabulary.

A projected human face would have tied the experience to human expressions. A circle gave us room to communicate through shape, colour, and movement: a nod, a shift of attention, a light bulb, or a mathematical symbol.

I designed the interaction states and their animations so that the robot’s voice and visual feedback worked together.

“We kept one glowing teal circle, then changed its shape and colour to respond to the question.”— My interaction design approach
THE ORIGINAL VISUAL LANGUAGE
Original Solbit animation communicating yes
01 / AFFIRMATIONA nod that says yes.
Original Solbit circle animation transforming to communicate an idea
02 / AN ANSWER“I have an idea.”
Original Solbit directional attention animation
03 / ATTENTION“I’m listening to you.”

Original interaction animations. Directional feedback used input from four microphones around the head.

03 / THE EXPRESSION LAB

Say something.
See it take shape.

Explore how one circle can communicate different meanings. Choose a prompt to see a response.

PROJECTED FACE / INTERACTIVE STUDYIDLE / READY TO LISTEN
TRY AN INTERACTION
THE RESPONSE

A quiet circle signals that Solbit is ready to interact.

Interactive recreations made for this portfolio. The maths, science, question-mark and fruit examples illustrate the visual language; original GIFs are shown above. No live AI or microphone access.

A MOMENT OF CARE / RETURNING TO CHARGE

“I’m a little tired.
I’ll take a rest.”

A practical need, expressed in language a child could understand.

When its battery ran low, Solbit would explain that it needed rest, then return to its charging dock. The words, animation, and movement helped make the transition from play to charging understandable.

Ready for one more question?

01 PLAYING02 TIME TO REST03 RECHARGING

Illustrative motion study of the described return-to-charge experience.

04 / THE THINKING BEYOND THE ROBOT

One companion.
A whole
ecosystem.

The child’s interaction was one part of a connected learning experience.

I designed the experience architecture across the robot, the parent portal, and teacher workflows. The goal was to connect lesson creation, parental oversight, and the child’s learning in one understandable journey.

  1. 01 / CREATE

    Teachers & experts

    Upload lesson documents and update learning content.

  2. 02 / REVIEW

    Parents

    Review lessons, send them to Solbit, and follow the child’s interactions.

  3. 03 / EXPERIENCE

    Children & Solbit

    Ask questions, show objects, and explore lessons through conversation.

  4. 04 / FOLLOW UP

    Adults stay involved

    Review information that needs attention and guide the next learning moment.

Explore the original ecosystem diagram

THE CONNECTED LEARNING JOURNEY

Teachers and creators prepare learning content.

Original project architecture linking creators, teachers, children and parents through the Guwoo network and master nodes to robots and mobile devices

Original project architecture illustration, with an illustrative animated walkthrough. Moving signals explain the experience flow; they do not represent live data. Swipe across to explore the complete diagram.

Vector reconstruction of Solbit’s side profile: rounded head, curved projection surface, camera in the sculpted neck, and three rounded supports with recessed wheelsVECTOR STUDY / SIDE PROFILE
DESIGNED AROUND A PHYSICAL PRESENCE

A body that
backs up
the conversation.

01 / A projected face
Expressive feedback without a touchscreen.
02 / A single neck camera
The initial model’s camera sat in the grey inset below the head.
03 / Three wheeled supports
A mobile base, with a head that could turn gently left and right.
04 / Spatial awareness
Floor mapping supported movement around the home and a return to the charging dock.

The vector study and interactive 3D model are approximations reconstructed from project photos, not production CAD. Sensor locations beyond those confirmed are not labelled.

05 / LEARNING WITH REAL CHILDREN

First, a touch.
Then, a
conversation.

Early prototypes went into classrooms around Shenzhen and Nanjing, with teachers involved in the sessions.

Children initially touched the robot’s face. Once Solbit began speaking and projecting responses, they started interacting through conversation. When the roughly 25-minute prototype session ended, children asked to keep playing.

This was an encouraging qualitative observation from early trials. It was a starting point for further development, rather than a measure of long-term learning outcomes.

Project photo collage showing children with recording equipment, children gathered around Solbit, and a group photograph with the robot
PROJECT RESEARCH ARCHIVE / ORIGINAL PHOTOGRAPHY
A CLOSER LOOK AT THE SESSIONS

People around
the prototype.

These project photographs show children and adults in a recording setting and children gathered around Solbit.

They add context to the early project work alongside the trial observations described here.

Two original project photographs: children and adults beside a studio microphone, and children watching a Solbit prototype
SESSION DETAILS / RECORDING SETTING & PROTOTYPE INTERACTION
WHAT WE OBSERVED

Curiosity became conversation.

Children explored the unfamiliar form through touch before discovering its spoken and animated responses.

WHAT NEEDED WORK

More time. More awareness.

Battery life limited early sessions. Teacher feedback informed plans for additional sensing, including side and rear cameras and more microphones.

WHAT I TOOK FORWARD

Make the system legible.

Attention, answers, uncertainty, and the need to recharge all deserve clear feedback—not just a successful response.

Trial observations and project history are based on my account. Additional sensors were proposed improvements, not features confirmed on the initial model.

THE THREAD THAT CONTINUES

Before AI became
an everyday conversation,
I was designing
how it could have one.

Solbit shaped the way I think about AI experiences: make the system’s state visible, give uncertainty a language, and design the whole journey around the people involved.

Those questions still guide my work today.

Let’s talk about human–AI interaction ← BACK TO SELECTED WORK
NEXT PROJECT / TELEVISION UX

A different kind
of conversation.

Explore TCL Smart TV

TCL Smart TV interface design