How I Used Claude AI to Write My Book (And Why It Wasn't What You Think)
Let me be clear from the start: Claude AI didn't write my book. I did.
But without Claude as my writing partner, "The Accidental Nonprofiteer" would still be sitting in a Google Doc as 5,000 words of unfinished potential, just like it had been for the past eight years.
When I tell people I used AI to help finish my book, I usually get one of two reactions: either "That's cheating!" or "Wow, so AI can just write books now?" Both responses miss what actually happened. The reality is more nuanced—and more interesting.
The Problem I Couldn't Solve Alone
By 2016, I had written what I thought was a halfway decent manuscript about nonprofit tech leadership. I had compelling stories from my decade helping build Stupid Cancer, frameworks that I knew worked in practice, and insights that could genuinely help other "accidental nonprofiteers" who found themselves building organizations without formal training.
But I was stuck. What I had read like a collection of really good blog posts, not a coherent book. I couldn't figure out how to create narrative flow between chapters, how to structure the content for maximum impact, or even whether I was writing a memoir or a manual.
Every time I opened the document, I felt overwhelmed. I'd make small edits, rearrange sections, and then close it again, no closer to having something publishable.
Sound familiar? If you've ever started a creative project and gotten lost in the middle, you know exactly what I'm talking about.
What AI Actually Did (And Didn't Do)
When I finally decided to try working with Claude AI, here's what I discovered:
What Claude Did:
Diagnosed structural problems I couldn't see after staring at the same content for years
Identified gaps in logic and flow between chapters
Suggested organizational frameworks that served the content better
Provided immediate feedback on drafts and revisions
Helped me see patterns across 5,000 words that I missed when reading linearly
Asked clarifying questions that forced me to articulate my actual message
Caught inconsistencies in tone and terminology across chapters
What Claude Didn't Do:
Write any original content for the book
Generate the stories, insights, or frameworks
Create the voice or perspective
Make editorial decisions about what to include or cut
Handle the emotional labor of actually finishing a long-term project
Think of it less like "AI wrote my book" and more like "AI helped me become a better editor of my own work."
The Collaboration Process
The actual process looked nothing like asking ChatGPT to "write a book about nonprofits." Instead, it was an iterative partnership that evolved over our collaborative sessions:
Phase 1: The Brutal Diagnosis
I uploaded my entire manuscript and asked Claude to identify what was missing. The response was both validating and devastating: I had good material, but I was trying to write two different books at the same time. I needed to choose between memoir and manual, then restructure everything to serve that choice.
This was feedback I couldn't get from reading my own work. When you're too close to a project, you can't see the forest for the trees.
Phase 2: Structural Surgery
With Claude's help, I reorganized the entire book around a simple principle: use personal stories to illustrate universal principles. Each chapter would start with a specific experience from my early career and building Stupid Cancer, then extract broader lessons that any organization could apply.
Claude was particularly good at saying things like: "This story about volunteer management in Chapter 8 would be more powerful if readers understood the email crisis from Chapter 3 first." Connections I couldn't see became obvious.
Phase 3: Voice and Flow
I'd write new sections or heavily revise existing ones, then share them with Claude for feedback on clarity, tone, and readability. Claude caught when I slipped into nonprofit jargon, when examples needed more context, or when transitions between ideas felt abrupt.
It was like having an editor who could read 5,000 words in seconds and immediately spot what was working and what wasn't.
Phase 4: The Publishing Push
When it came time to actually publish, Claude helped with everything from Kindle formatting to marketing copy to SEO optimization. Having a partner who could seamlessly shift from developmental editing to technical publishing guidance made the whole process less overwhelming.
What This Taught Me About AI and Creativity
AI doesn't replace human creativity—it amplifies it. The stories, insights, and frameworks in my book all came from my actual experience. But Claude helped me see how to organize and present that experience in ways that would serve readers better.
The quality of collaboration depends on the quality of input. Claude could only work with what I gave it. The better I got at asking specific questions and sharing focused excerpts, the more useful the feedback became.
AI is excellent at pattern recognition, good at structure, but you still need to bring the soul. Claude could spot inconsistencies across 11 chapters that I would never catch reading sequentially. But the voice, perspective, and emotional core of the book had to come from me.
Having an always-available writing partner changes everything. No scheduling conflicts, no guilt about "bothering" someone with half-formed ideas, no waiting for feedback. I could work on the book whenever inspiration struck.
The Ethics Question
Some writers worry that using AI diminishes the authenticity of their work. I understand the concern, but I think it misunderstands what's actually happening.
I didn't use AI to generate ideas or write content. I used it to become better at organizing my own ideas and presenting my own content. It's not fundamentally different from using a grammar checker, a developmental editor, or a writing coach—except that it's faster, cheaper, and available 24/7.
The book that resulted is more authentically "mine" than what I had before, not less. Claude helped me figure out what I was actually trying to say and say it more clearly.
For Other Writers Considering AI Partnership
If you're thinking about using AI in your writing process, here's what I learned:
Start with diagnosis, not generation. Don't ask AI to write for you. Ask it to help you understand what's not working in what you've already written.
Be specific about what you need. "Make this better" gets generic responses. "Does this transition work?" or "Is this example clear?" gets useful feedback.
Iterate rapidly. The real value comes from the back-and-forth. Share drafts early and often, use the feedback to guide your next revision, then share again.
Remember you're the author. AI can help you refine your ideas and improve your structure, but the insights and voice have to come from your actual experience and perspective.
Use it for what it's good at. AI excels at spotting patterns, identifying gaps, and suggesting organizational structures. It's less good at understanding nuance, emotional resonance, or what you're really trying to accomplish.
The Bottom Line
Eight years after I wrote the first draft, "The Accidental Nonprofiteer" is finally available on Kindle. It's a better book than what I had in 2016—clearer, more organized, and more useful to the people who need it.
AI didn't write my book. But it helped me write my book better.
And for a perfectionist who had been stuck in revision hell for nearly a decade, that made all the difference.
The technology served the creativity, not the other way around. Which, coincidentally, is exactly the principle I write about throughout the book when it comes to choosing tools that grow with your mission rather than distract from it.
Sometimes the best way to practice what you preach is to get out of your own way and ask for help—even if that help comes from an algorithm.
