Your Workflow Is Yours: Why Personalizing Your AI Process Matters More Than Following Someone Else's

There's no one right way to write with AI — and the sooner you stop looking for it, the sooner you'll find what actually works for you.

If you’ve spent any time in AI writing spaces recently, you’ve probably heard some version of this: “Here’s my exact workflow. Here are my exact tools. Here’s exactly how I do it — and you should do it exactly like this.”

And maybe you’ve tried to follow along. Maybe you’ve downloaded the same apps, copied the same prompt structures, and still found yourself staring at the screen wondering why it feels so wrong. Or maybe you haven’t even started yet, because the sheer number of options out there has made the whole thing feel impossible. Too many tools. Too many opinions. Too many people who seem to already have it all figured out.

I’m here to reassure you: you’re not behind. You’re not doing it wrong. And there is no one right way to use AI for writing. There never was.

The Problem Isn’t You. It’s the Noise.

When I talk to authors about AI — whether in the Future Fiction Academy, in our Facebook group, or just in conversation — I hear three versions of the same struggle over and over again.

The first is the copycat trap: someone finds a workflow they admire, tries to replicate it exactly, and can’t understand why it isn’t working for them. The second is the shame spiral: their process is working, but it looks nothing like what the person on YouTube is doing, so they assume they must be doing something wrong. And the third — the one that breaks my heart the most — is the defeat before the start. “I don’t even know where to begin. There are too many choices and I’ve never worked with this before.”

That last one is so common, and so understandable. The AI landscape right now is genuinely overwhelming. We are in a moment of enormous abundance — new tools, new models, new platforms, new opinions about all of them seemingly every week. And when you’re standing at the entrance to all of that with no map, it’s easy to just... not go in.

But the good news is that you don’t need a map of the whole territory. You just need to find your own path through it.

There Is No “Correct” AI Writing Stack

Let me be direct about something, because I think it bears repeating even though I’ve said it before: there is no universally correct way to use AI for writing. None. Anyone who tells you otherwise is selling something.

My husband works in technology. He loves AI. He uses it constantly in his work and finds it genuinely useful. He has also never once opened ChatGPT. Not even once. His favorite tools are ones that most people in my world have never heard of, and he has zero interest in the tools I use every day. And he is completely right to work that way, because his tools fit his brain and his problems.

That’s the thing about this moment in AI: we are all figuring it out at the same time, and the landscape is still chaotic enough that there is no settled consensus about the “right” tools. Some companies will succeed. Others will fold. Over time, the typical author’s AI stack will probably look more consistent than it does today — certain tools will prove themselves and stick around, and the noise will quiet down. But we’re not there yet. Right now, we all have to find what works for us.

Kimberly Gordon, who joined me on Episode 66 of Brave New Bookshelf, put it better than I ever could: “It’s not about the tool, it’s about you. If you know how you work and you are comfortable, just get used to talking with AI. That is the skill you need.”

That skill — learning to communicate with AI — is the one thing that’s genuinely portable. Every tool, every platform, every model is going to change. The ability to talk to AI, to explain what you need and iterate until you get it, goes with you everywhere.

My Own Journey: From Weaver to Architect and Back Again

I’m going to share something a little personal here, because I think it illustrates this perfectly.

At the Future Fiction Academy, we developed a set of AI Writer Archetypes to help authors figure out how they naturally want to work with AI. There are four: the Gardener, the Weaver, the Baker, and the Architect. (You can read more about them here.) When I first started using AI for my fiction, I was a natural Weaver. I’d write by instinct, follow where the story led, and weave in what the AI gave me alongside my own prose. It felt creative and alive and very much like me.

Then, for a long time, I tried to be an Architect. I built out my outlines and ideas first, had the AI generate the bulk of the first draft, and came in afterward to edit and refine. It was efficient. It looked impressive. It’s the kind of workflow that makes for great screenshots and testimonials.

I never really enjoyed it.

Not once did I sit down to an Architect session and think, yes, this is it. The process didn’t fit the way my brain works. I like to mull a story while I’m writing it — to discover things in the middle of a scene that change everything that comes after. The Architect process, as great as it is for many people, cut me off from the part of writing I actually love.

But the good news is that with AI, the mulling doesn’t have to take six weeks or six months anymore. I can write a novel in a much quicker timeframe and still work the way my brain naturally wants to work. I’m back to being a Weaver, and the work is so much better for it — and so am I.

The tools didn’t have to change. My willingness to stop doing it “the right way” and go back to doing it my way made all the difference.

Permission to Be Curious Instead of Correct

Dana Sacco came on Episode 65 and talked about choosing Antigravity over Claude because it matched her logical, systems-oriented brain. Kimberly Gordon uses coding applications and IDEs because she genuinely believes story is code, and because that mental model opens up creative possibilities for her that standard AI chat just doesn’t. Neither of them is doing it “right.” Both of them have found what works.

What strikes me most about Kimberly isn’t any specific tool in her stack — it’s her energy. She is endlessly curious, always experimenting, always excited to share what she’s found. She works with me at the FFA, and being around her is a genuine joy because she doesn’t approach AI as a problem to be solved or a system to be optimized. She approaches it as a conversation. That curiosity is the workflow.

You don’t need the perfect stack. You need enough curiosity to keep trying things until something clicks.

Okay, But Where Do You Actually Start?

If you’re one of the authors standing at the entrance feeling defeated before you’ve begun, I want to give you something concrete.

Start with the free tiers of ChatGPT, Claude, and Gemini. Just those three. The frontier models are the best place to get a feel for different AI personalities and tones — they’re the most capable, the most widely used, and the most talked about, which means there’s a huge amount of community knowledge about how to use them well. Spend a little time with each one and notice how they feel. Which one do you like talking to? Which one’s responses feel closest to what you’re looking for?

If you find yourself wanting a more flexible, workhorse option — especially one with lower costs and fewer restrictions — open an account at OpenRouter.ai. It gives you access to all the frontier models plus hundreds of open-source alternatives, and it’s a great place to explore once you’ve got a sense of what you’re looking for.

But don’t try to build an entire system on day one. Start with one small thing — a brainstorming session for a scene you’re stuck on, a summary of your current chapter, a list of possible character names. See how it feels. That’s it. That’s the whole first step.

The Only Thing I Can Promise You

I can’t tell you which tools will be best for you. I can’t tell you whether you’re a Weaver or an Architect, a Gardener or a Baker. I can’t tell you that the workflow that’s working beautifully for me right now will still be the thing I’m using a year from now — because honestly, it probably won’t be.

What I can tell you is this: if you stay open-minded, you will find the right tools for you. You will have to be willing to try things that don’t work out. You’ll have to be willing to walk away from an approach that looked good on paper but didn’t fit your brain. You’ll have to be willing to go back to something you abandoned, or try something you’d dismissed, because sometimes what wasn’t right at month three is exactly right at month twelve.

Your workflow is yours. It gets to look different from mine, different from Kimberly’s, different from every author you admire online. That’s not a failure. That’s the whole point.

---

I’d love to know: have you ever tried someone else’s AI workflow and found it just didn’t click? Or have you discovered something that surprised you by fitting your brain perfectly? Tell me in the comments — I read every one.