Behind the books · An author's note

The stories were always there.
AI helped me write them.

I've carried the world of Harriet Keel in my imagination for years. The people, the places, the humour, the situations and the stories I wanted to tell. What I lacked wasn't imagination. It was a practical way to turn all of that into finished books.

Generative AI changed that.

How the books came together

Before the first manuscript

Years in my head. Finally, on the page.

Like many people, I've had stories living in my head for a very long time. I could imagine conversations, see places, invent characters and picture the turns a story might take. But imagining a novel and sitting down to write it, sentence after sentence, are very different things.

I'm a designer by profession. I'm used to developing ideas, testing alternatives, making changes and refining a result until it works. When AI language models became available, I discovered that I could apply a similar process to storytelling.

Instead of leaving those imagined stories unwritten, I could start shaping them into chapters, challenge the results and keep working until there was a book to hold and share.

“AI didn't give me the desire to tell these stories. It gave me a way to finish telling them.”

The working method

Not one prompt. An ongoing creative process.

The manuscripts have evolved through many decisions and revisions. I use several large language models, compare their suggestions and decide what belongs in the finished story.

01
The story begins with me
I establish the ideas, settings, themes, conflicts, personalities and kind of story I want readers to experience.
02
Multiple AI collaborators
Different language models help me explore possible scenes, narrative structures, dialogue and alternative approaches. I compare, question and reject ideas as well as use them.
03
Characters change and grow
Names, ages, relationships, backstories and motivations are revised as the story develops. I make decisions about the world and how it fits together across books.
04
Plots are rebuilt
Scenes are moved, chapters rewritten, twists tested, continuity problems resolved and endings strengthened. A generated draft is a starting point for more work, not the finished product.
05
My editorial decisions
I decide what makes the characters believable, where the humour belongs, whether the pace works and what gets cut. I use AI for criticism too, but I make the final creative choices.
06
From novel to other formats
The same stories can be developed into audiobooks, visual material and television proposals, each requiring different approaches and further specialist work.

An open disclosure

What AI wrote, and what I created

I want to be clear: generative AI has written substantial portions of the prose and dialogue in my books, and has helped develop and revise the manuscripts. I would never claim to have personally typed every sentence.

My role has been to originate and direct the creative project, develop its world and characters, assess competing approaches and keep making editorial decisions. The books are human-directed and AI-assisted, not entirely hand-written novels.

That's a distinction worth making. Creative direction, writing individual passages and legal copyright ownership are not identical. As I work with professional editors, literary agents and producers, I'm committed to transparent disclosure and proper review of the relevant rights.

What comes next

The stories were always there.AI helped me write them.

Years in my head. Finally, on the page.

Not one prompt. An ongoing creative process.

The story begins with me

Multiple AI collaborators

Characters change and grow

Plots are rebuilt

My editorial decisions

From novel to other formats

What AI wrote, and what I created

More stories can find their way into the world.

The stories were always there.
AI helped me write them.