I firmly believe that we all need to decide for ourselves where we land on the issue of using AI for our nonfiction books. But I’ve noticed something tricky happening when I speak to nonfiction authors, particularly those who are in the early stages of their books. Very often, I’ll hear someone say, “Well, I don’t use generative AI to write, but I do use it to brainstorm.”
Or maybe, “I just use AI to organize my thoughts.”
I want to talk about that word just, because I think we may be underestimating how much happens during the brainstorming and organization stages of developing a book.
This is the third installment in our four-part series about AI in nonfiction writing. In the first installment, we talked about using AI to write a book, and in the second, we explored how AI is creeping into different stages of the publishing process. My goal with this series is to give you the information you need to make your own decisions about how you want to use large language models (LLMs) in your book development, or whether you want to use them at all.
Today, we’re going to focus specifically on using AI to brainstorm, organize existing content, and develop the ideas that will eventually become your nonfiction book. Because even though those activities might seem harmless, I think there are some consequences worth considering.
And if you’re hoping your book will become the foundation of a larger thought leadership career—through keynote speaking, workshops, social media, professional opportunities, or advocacy—this conversation is especially important for you.
Writing and Thinking Are More Connected Than You Might Realize
In the first episode of this series, I talked about how writing is really a series of choices. We choose which words to use, how to arrange them, what to emphasize, and how to communicate our ideas to readers.
I want to pull that idea into today’s conversation, because I think we often underestimate how interconnected writing and thinking really are.
It’s a bit of a misnomer to talk about using AI for thinking versus using AI for writing, because writing is, of course, thinking at its core.
Think about what happens when you sit down to journal. If you’re a regular journaler, as I am, I bet you’ve had the experience of beginning with a vague feeling or a problem you can’t quite articulate, and then finding yourself coming to a new understanding as you write.
You didn’t necessarily know what you thought when you picked up the pen. The act of writing helped you figure it out.
The same thing happens when we’re developing a nonfiction book. As we start putting our ideas into words, comparing them, questioning them, and arranging them, we begin to understand those ideas differently.
So when somebody tells me they’re using AI to brainstorm their book, I want to understand exactly what they mean. Are they using it to transcribe their thoughts? Maybe they’re asking it to suggest what their book should be about? Or feeding it years of existing content and asking it to identify themes or create an outline?
Those activities can have very different implications, and some of them involve handing over substantial parts of the thinking process.
The Temptation to Let AI Organize Your Existing Content
I want to focus on a particular scenario that I see fairly often with nonfiction authors.
Let’s say you’re somebody who has been creating content for years. Maybe you’ve been blogging, hosting a podcast, giving talks, appearing on other people’s podcasts, or publishing an email newsletter. Maybe you’ve been sharing your professional expertise on social media, or perhaps you’ve been keeping journals for decades.
You have a lot of material, and you know that somewhere in all of that material, there’s a book.
But the idea of sitting down and combing through years and years of content sounds absolutely miserable.
I get it! There are few things that sound more tedious than going through every blog post you’ve ever written, every podcast interview you’ve ever recorded, and every newsletter you’ve ever sent, trying to figure out how all of those ideas connect.
So it’s incredibly tempting to think, “Well, I’ll just upload everything into Claude or ChatGPT, ask it to identify the major themes, and have it put together an outline for me.”
On the surface, this sounds like an efficient way to tackle an unpleasant task.
In fact, we’ve developed a free Start Your Book Strategy Guide specifically for people who have accumulated years of existing content and want to turn it into a cohesive nonfiction book. The guide walks you through gathering that content, identifying connections between ideas, and developing an outline that will help you bring everything together.
But there’s a reason I recommend that you do this work yourself.
When you go through all of your existing content, you’re doing something much more valuable than taking inventory.
1. Using AI to Brainstorm Can Mean Missing Unexpected Connections
One of the things I love most about helping authors organize their existing content is watching them discover connections between ideas they had never previously thought of as related.
Imagine that you’ve been blogging regularly about five different topics. You probably think of those topics as five separate areas of expertise. You may have written dozens of posts about each one without ever sitting down to consider the intricate web connecting all five.
But when you go back through your work and look at those ideas together, something interesting starts happening.
You begin to see patterns.
You recognize that an idea you explored three years ago connects to something you’ve been thinking about recently. A story you haven’t remembered in ages suddenly comes back to you. You notice that two subjects you’ve always treated separately actually have a great deal to say to one another.
And sometimes, you discover something entirely new.
That last possibility is what really excites me.
Your Existing Ideas Can Lead You Somewhere New
Now, I know some of y’all are probably thinking, “But couldn’t I ask an LLM to identify those connections for me?”
Sure. You probably could.
An LLM can look backward through your existing work and identify recurring subjects, language, and relationships between ideas. It might do a perfectly reasonable job of telling you what you’ve already written about.
But there’s something important missing from that process: the forward-looking perspective that you bring as the expert.
When you read through your own work, you’re bringing everything you’ve experienced, learned, and thought about since you originally created that material. You’re also bringing your current questions, interests, frustrations, and ambitions.
So you might encounter an old blog post that suddenly reminds you of a story you haven’t told before. Or you might recognize an opportunity for a program, product, or service that you could offer your audience. You might even realize that the different topics you’ve been writing about for years actually contribute to one much bigger idea.
And friends, that’s where the gold comes in.
Sometimes, the process of organizing your existing work leads you to discover a topic that pulls all of your expertise into one cohesive concept. That concept might become the book you’re planning to write, but it can also send you forward into an entirely new phase of your career.
It could inform your next series of workshops, your next keynote, your next research project, or an entirely new way of talking about the work you’ve spent years doing.
If you’re a longtime content creator, I bet you’re hungry for that next step. You may have been sharing your expertise for so long that you’re ready to see where all of it can take you.
When you go through your work with your own eyeballs, making those connections yourself, you’re giving your brain an opportunity to discover something new.
An LLM can organize your material. But the point of the exercise is to see your material in a new way.
And if you skip the work of reading, remembering, comparing, and reflecting, you may also skip the discoveries that would have happened along the way.
This is also one of the reasons I’m such a fan of developing your book idea before you begin drafting. There’s a lot of intellectual work that happens before you ever write your first chapter, and it’s worth giving yourself the opportunity to do it.
2. AI-Generated Brainstorming Can Limit Your Book’s Structure
The second thing I want you to think about is structure.
If you’ve followed Page & Podium for any length of time, you know that I talk about structure constantly. I firmly believe that developing a strong outline is one of the most important things you can do before drafting a nonfiction book.
We’ve talked before about three different ways to outline a nonfiction book, including narrative structures, topical structures built around existing content, and persuasive structures.
But I want to zoom out for a moment and talk about what structure actually means.
At its core, structure is simply the way we define and communicate connections between the different points we want to make.
I think this is easiest to understand when we consider narrative nonfiction.
Structure Is About the Relationships Between Ideas
In a memoir or other narrative nonfiction book, we often rely on chronology to communicate relationships between events.
One thing happens, then another thing happens, then another. Because we’re moving through a sequence of events, the reader can usually understand how one moment relates to the next.
But we don’t always follow a perfectly chronological structure.
Sometimes we jump backward into a flashback. In other cases, we introduce an anecdote from outside the central timeline. Other times, we use an extended metaphor to help the reader understand something happening in the main narrative.
And when we do these things well, our readers come along with us.
Why?
Because we’re demonstrating a connection between those moments, even when the connection isn’t chronological.
The same principle applies to other types of nonfiction. A self-help book, a political advocacy book, or a book about professional leadership may have no chronological narrative at all. But the author is still responsible for helping the reader understand how each topic connects to the next.
That’s what we’re doing when we develop a book’s structure. We’re taking a 10,000-foot view of our ideas and arranging them so that the relationships between them become clear.
If you’re working on a narrative nonfiction book, our Memoir Method 7-Step Structure Sequence walks you through how to make those choices. And if you want a better understanding of what a finished outline should accomplish, our post on the anatomy of a memoir outline is a useful place to start.
But here’s where using AI to brainstorm your book can become particularly limiting.
The Most Interesting Connections Are Often the Least Obvious
In the first episode of this series, I talked about how LLMs generate language based on statistical patterns and probabilities.
That matters when we’re thinking about structure because the most probable connection between two ideas isn’t necessarily the most interesting one.
And if you’re writing a nonfiction book because you have a particular message you want to share with the world, I bet you’ve spent a lot of time thinking about how to make that message unique and true to you.
In my work with authors, some of the most exciting discoveries we make happen when we connect ideas that might not have seemed related at first.
That sometimes means connecting a childhood experience with a political encounter that happened decades later.
Or it looks like recognizing how an object from your childhood home reflects something you experienced in a workplace.
Maybe it’s discovering that two completely different academic disciplines offer insights that can help us understand the same phenomenon.
Those connections are often unusual precisely because they emerge from the author’s particular life, background, and expertise.
They’re also the kinds of connections I worry authors will miss when they outsource their brainstorming and structure development to an LLM.
The model may identify the most readily apparent relationships between ideas. But those are not necessarily the relationships that will make your book compelling, surprising, or uniquely yours.
Think about the books that have stayed with you long after you finished reading them. Very often, what makes them memorable is the author’s ability to show you something you hadn’t thought about before.
Maybe the author takes two seemingly unrelated experiences and demonstrates that they’re actually connected. Maybe they bring together ideas from different professional fields or historical moments and reveal an unexpected pattern.
That’s the kind of discovery that makes me excited about nonfiction writing.
And it’s one of the reasons I encourage authors to spend real time with their own material, even when doing so feels tedious.
Brainstorming and outlining are different activities, but they’re closely connected. The ideas you discover during brainstorming are what allow you to make the structural choices that will ultimately shape your reader’s experience.
You want those choices to reflect the connections that are most meaningful to you, not simply the most predictable relationships between the words you’ve already written.
There’s just no shortcut here, because the combing through is the work.
3. Outsourcing Your Thinking Can Make It Harder to Explain Your Own Ideas
Now, I want to talk about a third consequence that I think is especially important for thought leaders.
If you’re writing a book because you want to spread a message beyond the book itself, you’re probably hoping to spend a lot of time talking about your ideas.
Maybe you want to give keynote speeches or lead workshops. Maybe you’re hoping your book will create opportunities for podcast interviews, media appearances, or professional advancement.
Whatever that looks like, you’re going to need to be able to talk about your ideas in ways that aren’t perfectly scripted.
And I want to share a story that illustrates why this matters.
What Happened at an Academic Conference
A friend of mine recently told me about something that happened at an academic conference.
Now, if you’ve never been to an academic conference, these places are not necessarily as exciting as you might imagine. But they serve an important purpose. Researchers will often present brand-new work that hasn’t been published yet so they can discuss their findings with other scholars, answer questions, and strengthen their conclusions before publication.
That’s exactly what was happening at this conference.
A very well-respected scholar in my former field was presenting a piece of qualitative research. In this case, he and a couple of graduate research assistants had conducted interviews with a large group of participants.
Traditionally, analyzing this kind of research involves going through the interview transcripts, examining the language participants use, identifying similarities and differences, and looking for meaningful patterns.
I spent years doing this kind of academic work, and there’s a particular experience that comes with it.
You’re reading through one transcript, and somebody says something that catches your attention. Then you move into another transcript and notice that someone else is describing something similar, but in a different way.
And there’s this little flash of understanding.
You begin to recognize a pattern that you hadn’t seen before. You start thinking about what it might mean, how it connects to your existing knowledge, and whether it changes the way you understand the phenomenon you’re studying.
That process is a major part of qualitative analysis.
But in this particular case, the research team had decided to try something different. Because LLMs were relatively new and exciting when they developed their research plan, they decided to put their interview transcripts into ChatGPT and see what kinds of results it produced.
I understand the temptation. On the surface, it probably seemed like a reasonable way to make the analysis more efficient.
So when the professor stood up at the conference, he presented the findings the LLM had generated, along with the team’s interpretations based on their disciplinary knowledge.
And then the audience started asking questions.
This is a normal and important part of presenting academic research. Other scholars want to understand how you reached your conclusions, what you observed in the data, and how your findings connect to other work in the field.
But according to my friend, this professor couldn’t answer the questions.
To his credit, he was very forthright about what had happened. He acknowledged that because the LLM had produced the analysis, he hadn’t spent the time digging through the transcripts himself, as he had throughout the previous decades of his career.
He had the findings, but he didn’t have the same firsthand familiarity with the material that would have come from doing the analysis.
And that distinction matters.
Your Book May Be Only the Beginning
Now, think about how this applies to your nonfiction book.
Imagine that you’ve written a book on a topic you care about deeply. You want people to read it, of course, but you also want the book to create new opportunities for you to share your expertise.
You might be standing on a stage someday, talking about your work and taking questions from an audience. Or you might be on a podcast, having a conversation with a host who takes your ideas in a direction you weren’t expecting.
In those moments, you can’t simply repeat the text of your book.
You need to understand your material well enough to explain your reasoning, respond to challenges, and make new connections in real time.
And I believe that the process of developing and writing your book is a major part of how you build that ability.
When you’ve spent months reviewing your existing work, identifying connections, deciding what belongs in the book, developing your structure, drafting your chapters, and revising your ideas, you’
