LinkedIn Thought Leadership, Without the AI Slop
7 ways to work with AI so your writing still sounds like you.
Raise your hand if you’ve ever read a LinkedIn post that made you roll your eyes.
And AI, of course, has made it so that these posts are seemingly everywhere. It’s not singlehandedly responsible for the sharp drop in quality, but it does mean these posts are created en masse.
To be clear, when AI is deployed effectively, it’s a great tool that saves busy leaders’ time. The issue arises when the content lacks any human input, any tinkering or refining, any real intentionality. Your writing has to sound like you, because when people read AI slop, that erodes trust.
For start-up founders and leaders, trust is everything. Without it, you won’t reach the best talent, retain star employees, build investor confidence, engage the media, the list goes on.
So – how do you collaborate with AI to produce LinkedIn thought leadership that’s actually good? Here’s a list that is by no means exhaustive.
Set the stage
Before you even reach the point where you’re drafting a post, train AI on your voice. Share examples of speeches, previous posts, etc. that capture your voice and style well. Upload company narrative & messaging frameworks. Flag if there are any words/phrases you steer clear from, for compliance reasons or otherwise. Are there any peers or mentors whose style you really admire? Let AI know what you like about their thought leadership, and why.
Don’t shy away from your POV
Training AI on your tone solves the “sound like you” problem, but AI will still default to vague content and avoid taking a stance. Refine content as needed and work with AI to make sure thought leadership actually reflects your point of view. It’s ok to take a stand on an issue. If you change your mind about it later, you’re allowed to acknowledge that too.
Share clear feedback
Treat AI like a coworker whose work you are editing. Share clear feedback as you iterate together, and be sure to share the final, approved version with AI for its benefit. It’s an extra step that is the last thing on your mind when you’re racing to get content out the door, but once the dust has settled, take the time to tell AI why the final version worked. It’ll make future drafts much stronger.
Drop the formula
AI slop or not, LinkedIn posts can be so formulaic. Like, anecdote about something mundane -> transition that links the story to a career context. That doesn’t mean we need to stop sharing anecdotes – they are generally great to have. But the key to getting it right goes back to filtering AI-generated content for human taste and credibility. If the link between the anecdote and the career takeaway feels too exaggerated, then it might feel like AI slop because AI gravitates towards hyperbole.
Make the rhythm yours
Good writing has varied rhythm and cadence, whereas AI slop is repetitive. Read your draft out loud. If every sentence lands at the same beat, that’s usually AI’s fingerprints, not yours. Let a sentence run long and a little breathless if that’s genuinely how you’d say it, then follow it with something four words long. Rhythm engages readers.
Watch the vocabulary tells
Separate from rhythm and formula, because AI was trained on mountains of generic corporate-speak content, it reaches for certain words constantly: “delve,” “unlock,” “navigate,” “robust,” “game-changer,” the stacked em dash. Take the extra time to figure out what a better, more fitting word might be, and use that instead.
Make a judgement call
Not every AI habit or tell is a red flag. AI is trained on a mix of genuinely great writing and a lot of mediocre writing, which is exactly why it defaults to things like the rule of three so often. It’s a common tell, sure. But it’s also just a solid rhetorical device that’s been around a lot longer than ChatGPT has. If a list of three fits what you’re saying, keep it. The goal isn’t to strip out anything that resembles an AI pattern, it’s to keep what’s actually good and cut what’s just filler.
This isn’t an exhaustive list, just a starting point. I’ll be back with more of what I’m seeing (and what’s actually working) as this keeps evolving. If that’s useful, subscribe! This is exactly the kind of thing Startup Storytelling will keep digging into.
Originally published on Substack.