Deep-dive

Take a closer look at what went into delivering this solution.

Discovery

We began with a simple observation: many children aren’t connecting with traditional books anymore.

Literacy engagement is declining. Children who don’t see themselves reflected in stories, or who find the structure of books too rigid, are turning instead to games and apps where they control the narrative.

Through conversations with parents, educators, and the kids, we identified two overlapping problems:

1. Representation and relatability: many kids feel unseen in mainstream books.

2. Engagement style: children today expect interactive and responsive experiences.

This led to a clear insight:

What if kids could describe the stories they wanted, and have them come to life?

We iterated on this concept through:

• Ideation workshops with carers and kids

• Prototyping story prompts and testing how children described their ideal stories

• Testing existing story engines to understand the technical possibilities and limitations

The result was our MVP vision:

An app where children tell us who and what they want to read about, and the app delivers a personalised, safe, and engaging story.

Challenges

Throughout development, we faced several key challenges:

• Complexity vs. Simplicity

How do we give children the power to describe a story without overwhelming them? We developed a streamlined, child-friendly prompt system where kids pick or say a few simple ideas.

• Content Safety

Open AI models are powerful but unpredictable. We had to create guardrails so no matter what a child requested, the output remained safe, appropriate, and in line with educational goals.

• Technical Access

Building a flexible system that works without requiring specialist hardware or login accounts. We wanted immediate access on any device.

• Retention

Unlike games, stories needed to be compelling enough to bring children back without typical gamification tactics.

Use of AI

AI became the engine that unlocked custom storytelling at scale.

Here’s how we used it:

• Story Generation

We integrated AI models (OpenAI GPT) to generate stories based on user prompts. This allowed for dynamic, personalised narratives without requiring manual authoring for every request.

• Controlled Creativity

Rather than using raw AI output, we developed modular templates and prompt engineering to ensure story structure, tone, and length were consistent and age-appropriate.

• Image Creation

We leveraged AI-powered image generation (DALL·E) for character art and cover illustrations, giving each story a unique, engaging visual identity.