I need to tell you something about how my books get made, because the conversation around AI and writing has gotten loud enough that silence starts to look like hiding. I’m not hiding. I’m just a writer who hates revision and finally did something about it.
Here’s the short version: I write the stories. AI helps me edit them. That’s it. That’s the whole thing.
But the long version is more interesting, so let me tell it properly.
The Problem Was Never the Writing
I love writing stories. I love the messy, nonlinear, ADHD-fueled process of throwing scenes at a document out of order — dialogue first, then the action around it, then the emotional beats that make the whole thing breathe. My drafts are rough and human and imperfect, and I write them that way on purpose.
What I don’t love is revision.
Revision is slow. It’s visually intensive. It’s the kind of repetitive cognitive labor that makes my eyes glaze over — sometimes literally, given that I’m low vision. For years, the pattern was the same: I’d finish a draft, feel the rush of completion, and then stare down the mountain of editing work ahead of me and think, I could just write another book instead.
So I did. Over and over. I accumulated roughly ten completed manuscripts that had never been properly revised. Entire novels sitting in folders, waiting for an editing process that felt physically and mentally impossible.
The Insight
Editing is mostly pattern detection. Is the pacing dragging? That’s a pattern. Did a character’s eye color change between chapters? Pattern. Is a sentence doing too much work, or not enough? Pattern, pattern, pattern.
Machines are exceptionally good at pattern detection.
So I asked a simple question: what if a machine could find the problems, and I could decide how to fix them?
The Editorial Engine
What I built is, essentially, a small publishing house inside my laptop.
Instead of one AI doing everything badly, I created specialized editorial agents — each one focused on a different dimension of craft. One checks continuity. Another analyzes sentence rhythm. Another evaluates voice consistency, or sensory detail, or whether the humor is actually landing. They work the way a traditional publishing house does, with different editors bringing different expertise to the same manuscript.
Their feedback comes in a format borrowed from writing workshops: stars and wishes. Stars are things working well — moments where the prose is doing exactly what it should. Wishes are areas that could be sharper, clearer, more alive. They also produce structural notes about pacing, emotional arcs, and character consistency.
Here’s the critical part: these are suggestions, not automatic edits. The agents propose. I decide. No AI output becomes final prose without my hands on the keyboard, making the call about whether a change serves the story or doesn’t.
Trained on My Voice
Every editorial agent was trained on a corpus of over one million words of my own writing. Not generic internet text. My sentences, my rhythms, my particular way of building a scene.
This means when the system suggests a revision, it’s not imposing some default style. It’s pattern-matching against my own voice and offering me a sharper version of what I was already trying to say.
Sometimes the suggestions are so precise they make me laugh — little surgical strikes that tighten a sentence in exactly the way I would have, given another three hours of staring at it. That’s not the machine replacing me. That’s the machine knowing me well enough to save me time.
What This Changed
The editorial engine didn’t change my writing. It changed my relationship with revision.
For the first time, editing wasn’t a solitary visual slog. It was a guided process — a conversation between me and a system that understood what I was going for. The detection was automated. The decisions stayed mine.
This is how The Two-Flat Cats finally got its editorial pass. It’s how Bait and Switch got its revised edition. It’s how Single Player Co-Op — a memoir about DID, blindness, and survival — got the careful, sensitive revision it deserved. Manuscripts that had been waiting years were suddenly moving.
What This Isn’t
I want to be precise about the boundary, because it matters.
AI does not write my rough drafts. The messy, human, scene-out-of-order process is entirely mine. Every character voice, every plot turn, every emotional beat starts in my head and comes out through my fingers.
AI does not make final editorial decisions. It flags, suggests, and analyzes. I read, consider, and choose. The prose in the finished book is mine.
What AI does is bridge the gap between a completed draft and a polished manuscript — a gap that, for me, was previously impassable.
Why I’m Telling You This
I write furry fiction. I write queer fiction. I write in communities where trust between author and reader is the foundation of everything. I know that “AI” has become a loaded word, and I know why. The fear that machines will replace human creativity is real, and the flood of low-effort generated content makes that fear feel urgent.
But that’s not what’s happening here.
I didn’t hand my stories to a machine and ask it to write them for me. I built a machine that makes me a better editor of my own work. The stories are mine. The craft is mine. The tool is mine too — I designed and engineered every piece of the pipeline.
Writers have always built new tools. Typewriters. Word processors. Track changes. This is simply the next step: computational tools that help a writer see their own work more clearly.
I’m telling you this because I believe in transparency. Because you deserve to know how the books you read were made. And because the truth — that a blind, neurodivergent writer built an editorial engine to finally finish the novels she’d been carrying for years — is a better story than the one people are afraid of.
Austen Tucker is the author of The Two-Flat Cats, Bait and Switch, The Painted Cat, and Single Player Co-Op.
Leave a Reply