Writing with AI (and drafting killer LinkedIn posts)
Take a kernel of an idea and turn it into something you're damn proud of.
The challenge of creating something that truly resonates... truthfully, it’s a challenge that will exist no matter how advanced AI becomes.
But that doesn’t mean we can’t create with it. Creating with AI means playing a game of ping-pong where:
We take full onus of coming up with the best possible idea (“thinking is cheap”).
We feed ideas to AI, early and often.
We expect everything will need to be refined or reworked in some way.
We save our keystrokes for refining relentlessly.
By and large, it’s about saving keystrokes. With or without AI, ideas don’t climb from “meh” to “on message” in one step. Good writing never does.
In this post, we’ll go through the process of drafting a LinkedIn post from a kernel of an idea. But this post isn’t really about crafting LinkedIn posts. Or about murder. It’s about writing effectively with AI at our side.
Key Takeaways
When used effectively, AI transforms single ideas into full drafts, saving hundreds of keystrokes
Take a unique perspective, create a draft and then refine (or re-do) to make it better
Master AI as a drafting tool by mastering the process of coming up with strong ideas that have unique angles
Bulk writing isn’t a thing
Crafting a LinkedIn post might seem like a straightforward task. You could probably even bulk write them, yeah? Churning out multiple posts leads to diluted messaging, mixed insights, and a lot of undedicated energy.
But bulk-writing never results in quality. The essence of providing value is the ability to connect with a singular, focused idea. And painful as that might be, it makes sense doesn’t it? Humans can only read one thing at a time, so the same can only be true for writing.
The process for drafting killer LinkedIn posts with AI
Time to go. Let’s take an initial kernel of an idea and build it into a polished gem of a post. In this process, AI will rely on us for the best possible idea (topic and themes), and we’ll rely on it to help us work faster. Note: I chose to write a LinkedIn post for ease of digestion, but this works with content of any length.
It all starts with a kernel, slightly bigger than a burp but smaller than an idea. Where do kernels come from? For me, they’re the random thoughts that occur during the thinking portion of my days. They could be literally anything:
What would happen if Gmail went down for 24 hours?
The Cleveland Browns and their failed Deshaun Watson experiment
AI and the 2025 job market
Now I wouldn’t just write about literally anything, but you get the idea. For our purposes, we’ll use the following kernel:
Embracing change
Step 1: Pull the topic from your kernel
That kernel is enough to get started, but good content has a degree of specificity. Ever asked AI a bland question and get some bland writing in return? Enough said.
Our topic needs to be more specific. Here’s how to take the kernel and hit the right level of granularity:
Take a position: Rather than “embracing change”, try “embracing change is painful”
Flip the angle on its head: "Embracing change is less painful if…”
Counter common sense: “Embracing change is impossible”
For our topic, let’s go with:
Embracing change is impossible.
Step 2: Layer in themes (2-3 of them)
The themes dig deeper into the bucket of specificity. Our topic is more granular, but is still going to give AI too much leeway. Themes allow us to box in the canvas we want AI to work from, and good ones come when you apply a lot of the same tactics we used to scrutinize our topic.
The game is to detail 2-3 different tangents off your topic. You’ll be at your best when your themes are genuine, stringent, and maybe most importantly, relatable.
Our themes for “Embracing change is impossible”:
Theme 1: “Embracing change wrongly implies that change is either good (worth embracing) or bad (worth rejecting)”
Theme 2: “It’s impossible to not react adversely to change, it’s human nature”
Theme 3: “What we need to embrace are our choices and decisions”
Note: Scroll to the bottom of this post for a sub-prompt to help develop the themes for your topic.
Step 3: Run the prompt
The prompt below takes your topic and themes. Fill in your topic and themes and you’re off. Optionally, you can also take a little control over the tone of your post by changing “beautiful and heartfelt” to something else (“witty but pensive”, “terse and to the point”, “the written equivalent of jazz hands”).
Write me a beautiful and heartfelt LinkedIn post about: {TOPIC}.
TOPIC = “[TOPIC]”
The themes of this post should point to:
1. [THEME 1 - e.g. disruption through technology]
2. [THEME 2 - e.g. highlighting how AI can help find those disruptive opportunities]
And I now have my first draft. The key is: I shoot to do these steps 1-3 in two minutes or less.
Step 4: Iterate
With the right level of rigor applied to the topic and themes—you’ll be surprised at how decent that first draft can be. Never perfect, but decent.
With the first draft in hand, what happens next is up to your tendencies as a writer. But the goal is to iterate until it shines.
Two methods for iteration
You have to start by asking what you, as a reader, do and don’t like about the draft. From there, we will go into one of two directions:
Refining OR
Re-doing
Iterating through refinement
Refine when: The core of your draft works but it needs some polishing.
This could include:
adjusting the tone
tightening key elements
enhancing specific sections
For tonal adjustments, pull from your audience. If the draft is too formal, you might prompt AI to rewrite it with more warmth: "Make this feel like advice shared over coffee." On the flip side, if the tone is overly casual, you can request a more professional angle.
Exploring variations can also help tighten key elements. For instance, you might ask AI to experiment with calls-to-action (CTAs) or create multiple hooks:
"I used to resist change until I realized it’s the only constant in leadership."
"Every growth story starts with a chapter on change."
"Change isn’t the enemy—it’s the catalyst."
Alternatively, if you find specific sections that miss the mark, you can enhance those blocks or paragraphs separately.
Some writers might choose to abandon AI at this point and work somewhere else. But if you did keep working through AI, all of the above would be attacked via follow-up prompts.
Iterating by redoing
Redo when: The draft’s foundation feels shaky, or the original approach doesn’t land.
Starting fresh doesn’t mean discarding all progress—it means rethinking. If your draft feels fundamentally off, you can:
Tweak your prompt and re-run it (change themes, add restrictions like word count)
Use the accordion method to collapse the draft to its most key points
Explore alternative formats of your post
If it misses the mark entirely, don’t abandon the topic and theme. Instead, make some tweaks to the prompt and then re-submit (ChatGPT makes this very easy). You could alter the tone you’ve asked for, remove, add or reword your themes. You can even add restrictions like word count.
Another effective technique is the accordion method. Take the draft and condense it into two to four critical takeaways. Then expand these points into full paragraphs—this is where the ‘accordion’ part comes in.
You can even experiment with alternative formats to bring a fresh perspective. For instance, you might prompt AI to rewrite the content as a numbered list: "5 Ways Embracing Change Can Drive Personal Growth." Alternatively, a story-driven approach could add a compelling narrative element: "Last year, I..."
Like I said, creative ping-pong. Tweaking and honing your messaging until you’ve got something you’re in love with.
Tapping the greater potential of AI-powered creativity
But adopting AI into this process does more than make a better process. We’re actually building a whole new set of capabilities.
Determining your voice
Regardless of how you work, you can always further calibrate AI for future posts. For example, you can:
Take AI’s first draft of your post
Provide your final version
Have AI compare the two and describe the changes (or writing style)
Which means we effectively have a way to pre-prescribe some style change in your original prompt.
Rapid content pivots
AI makes it incredibly easy to adapt quickly. Need to shift tone or pivot from a personal story to an industry trend? AI enables us to change direction on a dime. Embrace it.
Brainstorm without limits
AI democratizes brainstorming, making it possible to explore countless variations in far less time. Backed by the "thinking is cheap" mindset, creating new versions with vastly different angles, tones, or structures is at least 100 times easier.
Posts I’ve written
It wouldn’t be fair if I didn’t provide any examples, would it? Here are two LinkedIn posts started with this process.
Write me a beautiful and heartfelt LinkedIn post about: {TOPIC}.
TOPIC = “the inaccuracy of the term expert beginners”
The themes of this post should point to:
- how the phrase “expert beginners” is a demeaning, misguided, stigmatic, and close-minded phrase
Write me a beautiful and heartfelt LinkedIn post about: {TOPIC}.
TOPIC = "apple and how they’ve changed the paradigm with earbuds that double as hearing aids"
The themes of this post should point to:
- disruption and highlight how AI can help find those disruptive opportunities
Sub-prompts
Whenever possible, I like to include sub-prompts that can be useful in support of a bigger picture prompt.
Sub-prompt 1: Compare first draft and your final version
Your goal is to familiarize yourself with my specific tone of voice and style. I have provided two pieces of content below:
1. your first draft
2. my final version
Please analyze the changes I made in my final draft and provide me a detailed description of my writing style that I can use in future prompts.
-- First Draft START --
[Paste writing sample 1 - 50-500 words]
-- First Draft 1 END --
-- Final Version START --
[Paste writing sample 2 - 50-500 words]
-- Final Version END --