Duckie
A playful way for children to keep asking why
Duckie turns a child’s spoken question into a short visual activity, then gives parents a simple record of what they explored.
- Team
- Celine Chang, David Dai, Kyle Bolor, Ricky Payares and Patricia Shu
- Scope
- Concept design, parent survey and interaction design
- Evidence
- 36 survey participants
A child's question does not always arrive when an adult can answer
Children ask questions while a parent is working, driving, caring for another child or trying to finish something else. The question matters, but the adult may not be ready to respond.
We looked at that short gap. Could a digital product help the child keep exploring for a few minutes without turning the experience into another endless video feed?
Duckie began as a voice-based learning concept. A child asks a question, hears an age-appropriate response and explores the answer through a short story, visual explanation or activity.
Duckie’s role is not to replace a parent or teacher. It is to keep one question alive until an adult can return to it.
Existing options trade context for availability
Parents and teachers can offer context but are not always available. Search and general chat tools can answer almost anything, but they are not built around a young child’s reading level or judgment. Video platforms are easy to use, yet the experience is usually passive and disconnected from the original question.
We saw room for something more focused: an experience led by the child’s curiosity, with enough structure for a parent to understand and manage it.
How might we help a young child explore one question independently for a short time while keeping the experience visible and understandable to a parent?
Interest opened the door, but trust shaped the design
We shared the early concept through a Qualtrics survey with 36 participants. Twenty-nine expressed interest in the idea.
That was an early signal to keep exploring. It was not evidence that children could use Duckie successfully or that the product would improve learning.
What changed in the concept
Multimodal guidance
Voice-first input with visual and audio instruction
Learning-first positioning
A question, explanation and practice loop
Parent visibility
A parent learning summary with clear controls
Age adaptation
Adjustable reading level and challenge depth
Offline continuity
Future downloadable, pre-reviewed activity packs
A child uses Duckie, but a parent decides whether it belongs
For the child
Clear guidance with little reading
- Speak instead of type.
- See one clear choice at a time.
- Move between listening, looking and doing.
- Receive feedback without feeling tested.
- Ask for help when something does not make sense.
For the parent
Visibility without constant monitoring
- Know what the child asked.
- Set time and content limits.
- Understand how the answer was created.
- Review interests without reading a full transcript.
- Remove saved voice or activity history.
We used Oliver, a curious six-year-old and early reader, as a behavioral scenario. His family background was not treated as a research finding.
One question, five short steps
The goal was to answer one question well, not keep the child engaged for as long as possible. We organized the experience as a short loop that returns the topic to a parent.
- 01
Ask
The child asks a question aloud.
- 02
Clarify
Duckie repeats the question so the child can confirm it.
- 03
Explore
The answer becomes a short story, visual explanation or activity.
- 04
Practice
The child tries one small challenge connected to the question.
- 05
Share
The topic appears in a parent summary for a later conversation.
The interaction stays simple, even when the question is not
- 01
Ask by voice
- User goal
- Let a child begin without spelling or typing.
- Design decision
- Duckie starts with one spoken question and keeps the screen visually quiet.
- Why it matters
- Voice lowers the reading demand, while a single prompt makes the next action easy to recognize.
- 02
Confirm what Duckie heard
- User goal
- Recover from speech-recognition errors before they shape the activity.
- Design decision
- Duckie repeats the question and offers three choices: continue, try again or ask an adult.
- Why it matters
- A child can correct the system without entering a complex error flow.
- 03
Match the activity to the question
- User goal
- Make the explanation easier to follow than a long chat response.
- Design decision
- Animal questions can become visual stories, while processes can use sequences and simple science ideas can use guided activities.
- Why it matters
- The format supports the subject instead of adding play after the answer as a reward.
- 04
Check understanding without a score
- User goal
- Give the child another way to work with the idea.
- Design decision
- The activity ends with one choice, match, order or short explanation.
- Why it matters
- The response can show when the explanation needs to be simpler without making the child feel tested.
- 05
Bring the parent back in
- User goal
- Help a parent continue the conversation without monitoring every moment.
- Design decision
- The summary records the question, topic, activity and whether the child asked for help.
- Why it matters
- It gives the parent useful context without turning the experience into a full transcript.
A narrow space between trusted help and always-available content
The first competitor table made Duckie look as if it combined every useful feature. The evidence did not support that claim. A more honest comparison shows the tradeoffs across familiar options.
| Option | What it offers | Where it stops |
|---|---|---|
| Parents and teachers | Context, judgment and trust | Not always available at the moment a question appears |
| General AI assistants | Flexible answers across many topics | Not designed around a young child's reading level or judgment |
| Companion apps | Engaging conversation | Learning structure and parent visibility vary |
| Children's video platforms | Easy visual access | Mostly passive and often disconnected from the original question |
| Duckie's proposed space | One question-led activity with a parent summary | A product direction that still needs child testing |
Duckie’s position is an opportunity-space hypothesis, not proof that it performs better than existing products.
A product for children needs stronger limits than an ordinary learning app
Before Duckie moved beyond a concept, the team would need to define how it handles data, unsafe questions, time limits and the difference between a product and a human relationship.
- 01
Parent consent before a child begins using the product
- 02
Minimal collection of voice and activity data
- 03
Simple controls for deleting recordings and history
- 04
Age-appropriate language, activities and feedback
- 05
No advertising or engagement loops designed to extend screen time
- 06
Parent-defined limits for time, topics and content
- 07
An ask-an-adult path for confusion or unsafe questions
- 08
Clear separation between Duckie and a human friend, teacher or caregiver
These are product requirements. They are not claims of COPPA compliance, clinical benefit, learning improvement or verified content safety.
The survey clarified the concept, not the outcome
What it clarified
What parents expected before trying Duckie
- More visual guidance and less reading
- A clearer learning purpose
- Parent visibility and controls
- Age-adjusted language and challenge
- An offline option for specific situations
What remains unknown
What still needs direct testing
- Whether children can complete the flow independently
- Whether answers are understandable and accurate
- Whether activities improve recall
- Whether parents trust the product after repeated use
- Whether any proposed pricing is realistic
Test the core loop before adding more adventures
- 01
Child sessions
Observe whether children can ask a question, confirm what Duckie heard, follow the activity, recover from an error and explain one thing they learned.
- 02
Parent interviews
Test whether Duckie's role feels clear, which controls matter, what belongs in the summary and what parents do not want stored.
- 03
Learning review
Work with an early-childhood learning specialist to review language, activity length and feedback before measuring learning outcomes.
The character was not the hardest part. Trust was.
Duckie started as a playful character that could answer children’s questions. The survey showed that parents needed to understand what Duckie was for, how it handled questions and where they remained in control.
That moved the concept away from a virtual best friend and toward a smaller role: helping a child explore one question until an adult can return to the conversation.
If I continued the project, I would spend less time adding adventures and more time testing the core loop with children, reviewing content quality and designing parent controls beside the child experience.
The goal is not to answer every question for a child. It is to make one moment of curiosity easier to continue.