AI as a Creative Tool: The Making of Rosie & Raven

When I began creating Rosie & Raven, I already knew what I wanted it to look like.

In fact, I had known for almost ten years.

The original inspiration for the book came from a vivid dream I had in 2016, in which I was holding an ancient illuminated manuscript filled with mysterious poetry and illustrations that somehow moved off the page. I woke up knowing exactly how the book had felt, but with absolutely no idea how to make something like it exist in the real world.

By the time I began Rosie & Raven in 2024, technology had finally started catching up with my imagination.

AI was still very new, and I initially began experimenting with it as a way to storyboard my ideas. But I quickly realized I could use it in a much more interesting way—as another tool in the collage process I had already been using as an artist for years.

I had taken photographs of illuminated manuscripts from the 1400s, whose intricate details and strange, magical worlds had always fascinated me. I began combining those references with contemporary botanical drawings, illustrations, photographs, textures, and my own visual direction to generate individual components for the book.

I would then take those pieces apart, alter them, combine them, repaint them, and assemble them into full-spread compositions.

And AI at that point was far from perfect.

Getting a character to stand in exactly the right position—or getting a hand to look remotely like a human hand—could be maddening. Sometimes the easiest solution was to photograph my own hands and use them as reference. Other times, I had a model pose in exactly the position I needed so I could build the image from there.

It became this strange and wonderful back-and-forth between old and new: 15th-century manuscripts, photographs, botanical drawings, digital collage, my own artwork, and this completely new technology that was evolving almost as quickly as I was learning how to use it.

I loved it.

I could suddenly experiment with ideas that would once have required an entire team of illustrators and production artists. AI wasn't deciding what the world looked like. It was helping me find the raw materials I needed to build a world that already existed in my imagination.

Then I had to make it move.

Once the illustrations were finished, I used the same process I had developed over nearly a decade of creating my augmented reality artwork.

The animations themselves were built by hand in After Effects, often frame by frame and layer by layer. A 30-second animation could take me an entire month to complete.

But here again, AI gave me new possibilities.

I could generate small pieces of movement—a fairy flying through the air, for example, or the petals of a rose slowly opening—and then incorporate those elements into the much larger animations I was constructing.

Before these tools existed, creating something at this scale would likely have required an entire animation studio. Suddenly, as one artist sitting at my computer, I could do things I genuinely never imagined I'd be able to do on my own.

That, to me, is what was so exciting about AI. It didn't replace the creative process. It expanded what I was capable of creating.

And then I decided to make things even more complicated.

The prototype of Rosie & Raven relied on an existing augmented reality app to activate the animations. It worked—but it didn't feel magical enough.

I didn't want a child or parent to have to download an app, create an account, or navigate someone else's platform before experiencing the book. I wanted them to scan once and simply watch the pages come alive.

So I spent another six months working closely with an AR company to develop a custom platform specifically for Rosie & Raven. The final experience is completely web-based: no app download required, and one scan unlocks all of the animated pages throughout the book.

Of course, by then I had another idea.

What if the book had its own music?

I wrote Rosie & Raven in rhyme. One night, while searching through royalty-free music for the animations, I started absentmindedly singing some of the words of the book over the tracks.

And suddenly I thought: Wait. Are these songs?

I had sung a cappella in college, so eventually another slightly terrifying thought followed: Why don't I sing them myself?

I hired a composer, and together we turned those little melodies into six original songs using my own voice. For the final animation, I even convinced my dad to play violin.

What began as a children's book had somehow become an illustrated story, an animation project, an augmented reality experience, a musical—and one enormous creative experiment.

The tool isn't the artist.

Yes, AI is part of how I made Rosie & Raven. And I'm genuinely excited about that.

But it is one tool among many.

Creating the finished book drew on skills I had accumulated over more than 20 years as an artist, art director, creative director, designer, animator, and storyteller. It involved writing and rewriting, photography, collage, illustration, animation, music, singing, technology development, endless experimentation—and a tremendous amount of trial and error.

The entire journey took nearly two years.

What excites me most about AI isn't the idea that technology can create things for us. It's that it can give individual artists access to creative possibilities that once belonged only to large studios with large teams and large budgets.

For the first time, I had tools expansive enough to help me build the strange little world I had seen in a dream almost a decade earlier.

And somehow, piece by piece, I finally got to make it real.