SpellBoard

A small learning experiment for my son

Role

UI/UX Designer

Industry

Education

Duration

2 Weeks

Why I built this

SpellBoard is a small interaction experiment I created while helping my son practice spelling at home. English is not my first language, and raising children in English has made me very aware of how confusing spelling and phonics can feel while a child's confidence is still developing.

I wanted to explore what spelling practice could feel like if it were interactive, playful, and inviting instead of repetitive. Parents enter spelling words, children choose a themed world, and letters become part of a small adventure where they drag and place tiles to build each word.

The project was prototyped using Figma Make, which allowed me to quickly generate interface structures, explore visual directions, and iterate on interactions. From there, I refined layouts, hierarchy, and components manually using my design judgment, treating AI as an accelerator for exploration, not a replacement for craft.

Problems I ran into and how I solved them

This project went through many iterations. Several aspects of the experience needed refinement before the interface felt calm, intuitive, and cohesive.


CHALLENGE 01

Refining the Interface

The AI-generated starting point was cluttered, background layers competed with the interaction area and the layout lacked clear hierarchy. I simplified the visual system by reducing background noise, refining container shapes and spacing, and improving the grid. The goal: a calm, focused interface where the spelling interaction takes center stage.


CHALLENGE 02

Improving the Voice

Pronunciation feedback is critical for spelling practice. Basic text-to-speech sounded robotic and undermined the warmth of the experience. I researched alternatives and integrated ElevenLabs voice generation, which produces significantly more natural speech, improving both clarity and the sense that care had been put into the game.


CHALLENGE 03

Interactions for Kids

Early letter tiles and drop zones were too small and cramped for children to use comfortably on a touchscreen. Through several iterations I refined tile size, spacing, drop zones, and highlight states to make the mechanic immediately understandable, without needing instructions.


CHALLENGE 04

Motion and Feedback

To make the experience feel alive, I added small animations: feedback when letters are dragged, when tiles snap into place, and when progress advances along the adventure map. I used AI to explore timing possibilities, then refined behavior manually so everything felt purposeful, not decorative.


I used AI tools to generate starting points and explore possibilities, every decision about what stayed, what changed, and what felt right was mine.

REFLECTION

What this project was really about

SpellBoard started as a small way to help my son practice spelling. It became a design exploration about how learning tools can feel playful, calm, and confidence-building instead of repetitive or intimidating.

As someone who learned English later in life, designing a spelling experience in English feels personal. I know what it is like to feel unsure about how a word is spelled or pronounced, and I wanted the game to carry none of that anxiety.

This project also became an opportunity to explore how AI can support the design process, not by replacing design thinking, but by accelerating exploration and allowing more time for refinement and decision-making.

AI tools are most useful as accelerators for exploration, not final answers, the design judgment still matters.

  • For children's interfaces, interaction affordances need to be substantially larger and clearer than adult defaults.

  • Voice quality has an outsized impact on perceived quality in learning tools, a robotic voice undermines the whole experience.

  • A calm visual system requires active restraint, the tendency of generated interfaces is toward visual busyness.

  • Personal motivation makes for better design briefs. Designing for my son gave me a clear, honest user to design for.


Want to try it?

You can explore SpellBoard yourself and see how the experience works.

Enter a few spelling words, choose a world, and start the adventure.

👉 Try SpellBoard

REFLECTION

What this project was really about

SpellBoard started as a small way to help my son practice spelling. It became a design exploration about how learning tools can feel playful, calm, and confidence-building instead of repetitive or intimidating.

As someone who learned English later in life, designing a spelling experience in English feels personal. I know what it is like to feel unsure about how a word is spelled or pronounced, and I wanted the game to carry none of that anxiety.

This project also became an opportunity to explore how AI can support the design process, not by replacing design thinking, but by accelerating exploration and allowing more time for refinement and decision-making.

AI tools are most useful as accelerators for exploration, not final answers, the design judgment still matters.

  • For children's interfaces, interaction affordances need to be substantially larger and clearer than adult defaults.

  • Voice quality has an outsized impact on perceived quality in learning tools, a robotic voice undermines the whole experience.

  • A calm visual system requires active restraint, the tendency of generated interfaces is toward visual busyness.

  • Personal motivation makes for better design briefs. Designing for my son gave me a clear, honest user to design for.


Want to try it?

You can explore SpellBoard yourself and see how the experience works.

Enter a few spelling words, choose a world, and start the adventure.

👉 Try SpellBoard

Other projects

Josefina Poventud 2026

Josefina Poventud 2026

Josefina Poventud 2026