In partnership with
There is a particular kind of frustration that comes from having a finished project sitting right in front of you and still not being able to write about it. You know what you made. You know what you used. You could explain it out loud to anyone who walked in the door. But the blank page? The blank page just sits there.
For me, that gap between knowing what I want to say and being able to get it onto the page is not just a creative problem. Nearly twenty years of chronic migraines have made some days harder than others for the kind of sustained focus that writing requires. There were projects I finished and never wrote about. Content I meant to create that just quietly disappeared into the pile of things I did not get done.
AI changed that for me. Not because it writes for me, but because it gives me somewhere to start.
It Is Not Going to Hand You the Finished Thing
I want to be honest about what AI actually does, because the expectation gap is the reason most people either dismiss it or give up on it.
For example, if I were to set a lump of clay in front of AI and ask for a sculptural masterpiece, I’m not going to get Michelangelo's David. I am going to get something roughly human shaped. And then I have to say, I want to define the hands differently, I want more structure across the shoulders, and what are we doing about the fig leaf situation. That back and forth is how the thing gets built.
What AI gives you first is a shape. Something to react to. And if what it gives you looks finished, it probably looks like everyone else's version of finished. Generic. Smooth in a way that does not sound like any particular person.
The goal is not to take what it gives you and publish it. The goal is to use it as a starting point for a conversation, and let that conversation pull out the thing you actually wanted to say.
The Brain Dump
Here is what I actually do.
I open a chat window and I start talking. Not in an organized way. I dump in all of my thoughts: the project I finished, the products I used, the ones I do not carry anymore, the ones I want to sell more of, the vague idea I have about what I want the article to accomplish. It comes out messy and half formed and in no particular order. Sometimes I contradict myself. Sometimes I realize halfway through a sentence that I am not sure what I actually think.
Today I want to write a blog post and promotional email about a vintage doll case makeover that I did. I've got lots of pictures and video. I used chalk paint, a clay paint. It is a DIY paint by Debbie's Design Diary. I don't carry the paint line anymore, so I can post links to someone else's paint. Maybe I'll send them to my friend Wendy.
In the meantime, I also used IOD's Paradise Paint Inlay. I don't think I have any of those left in stock, so I want to write the article more generically about using paint inlays. I also used transfers, and the key that made it nice for me, I think, was the layering of different tools.
Then I ask AI to ask me questions. This is a key step.
Anyone can paste a prompt and get an article back. The part that keeps this from reading like a generic AI article happens before any writing starts. I ask it to ask me questions.
Can you ask questions to help me make this informational and to sell some more transfers and paint inlays, please? I want it to be good for SEO because it's going to sit on the blog. Ask questions. I can share pictures. What would be helpful to you for writing this?
What comes back is a set of questions I actually have to think about. What was the moment I realized this was worth writing? What would I want a reader to walk away knowing? What did I learn that surprised me? My answers to those questions are where my voice lives. AI can write a sentence. It cannot have my experience. The interview is how I put my experience into the article before the article exists.
In the session that produced this blog post, I came in knowing I had a project to write about. A vintage doll trunk I found for five dollars, painted, and sold. I had photos, some video, and a general sense that I wanted to promote my transfers and paint inlays. What I did not have was a clear angle, an SEO (Search Engine Optimization) strategy, or any idea how long the piece should be.
By the time we had talked through the questions, I had all of that. Not because AI handed it to me but because AI asking the questions forced me to figure out what I actually wanted.
Where It Falls Short
I want to be honest about this part too, because the shortcomings are real and worth knowing going in.
AI is not good at product specifics. In that same session, it kept referring to the Brocante transfer in a way that made it sound like Brocante was a type of transfer rather than the name of a specific product. I had to go back through the draft and fix that everywhere. It is a small thing, but it matters when accuracy is part of how you build trust with your audience.
One thing I have learned to watch for: dashes. AI uses them constantly, and they are now so associated with AI-generated content that people have made memes about it. I have saved a note in the AI’s memory telling it not to use them in my writing. I still catch them.
It is the most reliable sign that something came out of an AI and went straight to publish without a human reading it first. If you take nothing else from this post, take this: read your output before you post it. Not for spelling. For the things that make it sound like a robot had feelings about your doll trunk.Repeated rewrites get muddy fast. The more times you ask AI to take another pass at the same piece, the more it starts saying the same thing three different ways in the same paragraph. It adds synonyms, splits sentences, pads the whole thing. Even when you tell it specifically what is not working, it still loses clarity after enough rounds.
This is not a user error. It is just how AI works. At some point the tool has exhausted what it can do with that material, and the answer is to step away from the rewrite loop entirely. Paste the draft somewhere else, read it with fresh eyes, and make the final changes yourself. That last mile is yours anyway.Finally, AI cannot know what a project felt like to make. It cannot tell you what surprised you, what went wrong, what made you step back and think yes, that is it. That is yours. That is the part you have to bring.
This Took Time to Develop
One more honest note about this process. The version I am describing did not happen in a single session. I had to teach AI how I like to write and AI had to teach me its limitations.
AI writes better in my voice now than it did when I started, because I have shown it what my voice looks like over time. I have shared pieces of my writing, pushed back when something did not sound right, and redirected when it went in a direction I did not want. That history matters. If you are just starting out, the early sessions will feel clunkier. That is normal and it gets better.
The goal is to build a working relationship with a tool, the same way you would learn any other tool you use in your business. It takes some time before it feels natural. It is worth the time.
If You Want to See Exactly How This Went
The blog post that came out of this process is linked here:
If you want to see the full unfiltered and unedited conversation that produced it, every question, every redirect, every moment where I pushed back or changed direction, you can find the complete chat transcript here.
It is not a polished demonstration. It is the actual thing.
Talk to your AI tools the way you'd talk to a colleague.
You don't send a colleague a three-word brief. You explain the context, the constraints, what you've already tried. But typing all that into ChatGPT takes forever — so you don't.
Wispr Flow lets you speak your prompts instead. Talk through your thinking naturally and get clean, paste-ready text. No filler words. No cleanup. Just detailed prompts that actually get you useful answers on the first try.
Millions of users worldwide. Works system-wide on Mac, Windows, and iPhone.