The Challenge of Revising a First Draft
As I finished up the first draft of my novel a few weeks ago, I looked at the reams of notes I had made along the way—ideas to revise, plotlines to adjust, and characters to refine.
As countless writers have noted, the first draft is not about creating a finished product. It’s about discovering what the story is truly about. It wasn’t until I reached the end that I understood how some of it was supposed to begin. That’s the fun part—uncovering and shaping the story.
The not-so-fun part? Figuring out how to piece all of these notes together in a logical way so my second draft would be closer to the story I envisioned in my mind’s eye.
That’s where AI came in.
AI as a Creative Partner, Not a Replacement
I use AI regularly in my day-to-day workflow for content production at a fintech company and for writing in a marketing firm. The key to remember is that AI won’t replace you—because the human element of writing is still something that cannot be fully replicated. Instead, it can elevate your work by streamlining processes.
This was a lengthy but worthwhile process I engaged in. In the end, I came away with a fully developed outline and a rough idea of how books two and three in my trilogy would unfold.
Why AI Struggled with My First Draft
At first, I tried having ChatGPT analyze my first draft. My hope was that it could help me implement my new ideas into the existing structure and mold a revised outline. Unfortunately, it struggled with reading and retaining the roughly 92,000 words and often hallucinated in its analysis.
For example, it claimed that 0% of the book contained dialogue. I wasn’t surprised, though. This task was simply beyond what AI is currently capable of handling.
I considered breaking up the novel into smaller PDFs to feed into GPT, but I ultimately decided against it. Instead, I produced a very rough sketch of my revised outline based on my notes and the current plot structure. This became a springboard—I fed it into GPT and asked it to flesh out a more complete outline for me to work with.
Iterating on the Outline with AI
From there, I engaged in an extensive creative brainstorming process with GPT. I used various prompts to analyze my story through different structure archetypes and identify ways to align my plot more closely with those frameworks.
This often led to exploratory rabbit holes—sometimes GPT suggested half-baked ideas or ones that didn’t align with the story’s current structure. But rather than discarding these suggestions outright, I used them as opportunities. I would refine GPT’s ideas by asking clarifying questions, breaking out singular elements, or revisiting my notes and telling it, “Actually, here’s what we’re doing in this part of the story—how does this align with the problem you identified? Should we adjust the approach?”
At times, GPT suggested ideas I wouldn’t have considered on my own. I certainly didn’t use all of them—AI is programmed to keep suggesting improvements, sometimes going too far in an effort to appear helpful. Recognizing when to ignore certain suggestions was key.
This was where having written the first draft truly helped me. I had a much deeper understanding of my story than when I began writing, allowing me to filter out what did and didn’t serve my vision.
Organizing the Insights
To ensure I captured the most useful ideas, I frequently asked GPT to summarize our discussions. Given the amount of back-and-forth, this helped me keep track of key decisions. If the summaries were missing details or felt off, I’d probe further. If they were solid, I’d copy them into my working folder.
By the end, this folder had grown significantly. I had separate documents for each character profile, various backstories, settings, and major plot points—resources I could easily pull from when revising my second draft.
Finalizing the Outline
This weekend, I sat down with my current outline and went through all of these standalone documents. I pulled key plot points and integrated them into the outline, creating a more complete picture of how everything would unfold. Finally, I asked GPT to analyze this updated version.
At this stage, AI provided a few more solid ideas to fill in gaps, though some suggestions became repetitive or contradictory. This was another reminder that AI is most useful as a tool for refinement rather than an infallible guide.
The Role of AI in My Writing Process Moving Forward
In all, this process required a lot of mental effort and some manual work, but AI helped expedite much of the heavy lifting in organization and brainstorming. It served as both a creative partner and an administrative assistant.
That said, I won’t use AI to write my book. That would defeat the purpose for me. Writing this novel is a lifelong goal, and I want to tell the story myself. But I will continue to use AI to help with research and brainstorming during my second draft to ensure the story turns out as I envisioned.
AI can be a powerful companion in the writing process—but ultimately, the story is yours to tell.