← Video Library · Book Publishing
The AI-Assisted Writing Process: Refining Drafts for Publication
Published April 1, 2024 · 171 views on YouTube
Key takeaways
- An AI first draft is a starting point, not a finished manuscript — it still needs a full editing pass.
- Feed the AI an outline and a subject matter expert's knowledge base so it fills gaps rather than guessing.
- A professional editor reviews the draft for tone and flags spots that need more context or a different direction.
- Skipping human review and publishing an AI draft hands-off is described as rolling the dice on what actually goes out.
- Great authors gather information from many sources, filter out noise, and connect the resulting ideas into chapters.
Is an AI-generated first draft ready to publish as-is?
No. The video is direct on this point: it is unrealistic to expect an AI first draft to come out “perfect and publish and crystal clear” without guidance. Publishing it hands-off means, in the speaker’s words, “rolling the dice with what’s actually getting put out there.”
What does the AI-assisted writing process actually look like?
The process described has three main stages: start with an outline and a knowledge base pulled from a subject matter expert, have the AI use that material to produce a first draft that fills in the gaps, then hand that draft to a professional editor for a refinement pass.
- Gather an outline and a knowledge base of context from a subject matter expert.
- Let the AI generate a first draft using that outline and knowledge base to fill in the gaps.
- Hand the draft to a professional editor.
- The editor analyzes the draft for tone and flags where more context is needed or where a chapter drifted off direction.
- Revise the chapter based on that feedback.
Why does starting with an outline and expert knowledge base matter?
Because it narrows what the AI is working from. The video contrasts two scenarios: wanting to write “on the subject matter of board games for aspiring entrepreneurs” with nothing else to go on, versus having a draft, an outline, and an entire knowledge base of context for the AI to pull from. The first could go in a number of different directions; the second gets the output much closer to what’s actually needed.
What does a professional editor add after the AI draft?
The editor analyzes the draft for tone and can say, for example, that a section needs a little more context, or that a particular chapter was “a little bit off in this direction” and needs refining. That editorial pass is what turns a serviceable AI-assisted draft into something closer to publication-ready.
What separates a great author in this process?
According to the video, it comes down to the ability to pull information from many different sources, sift through the noise, pull out the useful nuggets, bring those nuggets together, and make the connections inside the chapters that help readers connect the dots themselves.
Full video transcript
That’s right, and then it’s going to get you that first draft, but of course it’s going to go through another round of editing, and I think it’s unrealistic to think that it’s going to be perfect and publish and crystal clear, hands off, like without some guidance. If you’re doing that, then you’re kind of rolling the dice with what’s actually getting put out there.
But again, it’s like you take that first draft that was taking the outline from the expert and the knowledge base from the expert, and then filling in the gaps, and then you hand that to a professional editor. Yeah, they can analyze it for tone. They say, hey, we need a little bit more context here, we need a little bit more here, we need a little bit more here. Hey, can we refine this chapter? It was a little bit off in this direction. It’s like, okay, but we’re a lot closer than we would be if just, like, I want to write on the subject matter of, you know, board games for aspiring entrepreneurs, and like that’s it. It could go a number of different directions.
Versus, if I had a draft, if I had an outline, if I had an entire knowledge base of context that I wanted it to pull from, then the more of that I can gather forth is first, and the organization of that content is, to me, what makes great authors great. Great authors are being able to pull all this information from all these different sources, sift through a lot of the noise, pull out the nuggets, bring those nuggets together, make the connections inside of the chapters, and helping people connect the dots like that in itself.