LinkedIn Posts Have Gotten Weird — And Your Readers Can Tell
LinkedIn has gotten complicated with all the AI slop flying around. Mid-2023 is when I first really clocked it. The polished posts — the ones with perfect hooks and three tidy bullet lessons — were dying. Zero comments. Tumbleweeds. Then someone I followed posted about their coffee machine breaking twenty minutes before a client Zoom, and it pulled 300 comments by noon. That was a Tuesday.
Your readers have developed a sixth sense for this stuff. They can’t articulate it. They just feel something’s off, the same way you feel when a smile doesn’t reach someone’s eyes. And once they feel it, you’ve lost them — credibility gone, connection gone, scroll.
The patterns are painfully predictable. Hook framed as a question or a hot take. Three structured takeaways. Soft call to action at the bottom. Every sentence landing with purpose. No friction. No weird tangent. No moment where the writer sounds genuinely unsure. That’s the tell. Real humans backtrack. We overshare. We throw in a detail that barely fits the narrative — because it actually happened and our brain won’t let us leave it out.
Probably should have opened with this section, honestly. But here’s the thing that really stings: AI models trained on LinkedIn data learned from the posts that performed best in 2019 and 2020. Same beats. Same architecture. So when you use AI to draft your post, you’re not just sounding like AI — you’re sounding like AI doing a haunted impression of 2019 LinkedIn culture. Hustle quotes and gratitude lists and all.
The fake vulnerability is its own disaster. An AI-written “vulnerable” moment reads like someone performing vulnerability in a corporate training video. “I failed 47 times before I succeeded.” Oddly specific number. Unnaturally rounded. Followed by a perfectly constructed three-point lesson with zero loose ends. Real vulnerability doesn’t resolve cleanly. It involves specific people, specific dates, sometimes dialogue you still think about at 2am. That’s what makes messy human writing endearing to us readers.
The Six Phrases That Instantly Kill Your Credibility
These are the tells. Not subtle ones — obvious ones. When readers see these phrases, something in their brain just flags it and moves on.
- “In today’s fast-paced world” — This phrase appears in approximately 8 million LinkedIn posts a week. I’m apparently not exaggerating by much. It does no actual work. Replace it with something that happened: “Last Tuesday, I watched three meetings get canceled back to back…” Start with the moment. Skip the stage-setting entirely.
- “I am thrilled to announce” — You’re probably not thrilled. You’re probably relieved, or quietly proud, or terrified and pretending otherwise. Say that. “Thrilled” only shows up when something is trying to sound enthusiastic without any specificity behind it. Don’t make my mistake — I used it in a post once and my old manager texted me asking if I was okay.
- “Here is what I learned” — This is the AI transition phrase. Just tell the story. Let the lesson land on its own. Your reader will get there. You don’t need to announce that the lesson is incoming, like a flight attendant preparing the cabin for landing.
- “Game-changer” or “paradigm shift” — Drained of all meaning. If something genuinely changed how you work, describe what changed. Before and after. Actual numbers. Real impact. Skip the hype word entirely — it’s doing the opposite of what you want.
- “Leverage” (as a verb) — Real people “use” things. They “try” things. They “steal” things in the good way, or “borrow” ideas, or “gut-check” approaches. Pick the verb that describes what actually happened at your desk at 3pm on a Wednesday.
- “Delve into” or “unpack” — These make you sound like you’re hosting a podcast from 2015. Just say what you’re doing: “I want to talk about why this keeps happening” or “Let me explain what went wrong with our Q3 launch.”
Before and After — Same Post Rewritten to Sound Human
Example 1: The Career Update
BEFORE (AI Draft):
“Excited to announce that I’m joining Acme Corp as VP of Operations. This is a game-changer for my career, and I’m thrilled to leverage my experience in process optimization to drive impact. In today’s fast-paced world, operational excellence is more important than ever. I’m grateful for the mentorship I’ve received along the way. Here’s what I learned: 1) Consistency matters 2) Always keep learning 3) Your network is everything. Let’s connect and grow together.”
AFTER (Human Version):
“I’m joining Acme Corp next month as VP of Operations. Honestly, I’m nervous. Their ops team is 40 people — I’ve never managed more than 18. But they’ve got something genuinely strange happening with their inventory system, the kind of problem I’ve spent three years thinking about, so here we are. Weirdly grateful to my last boss for telling me no back in 2019. That one refusal sent me down a rabbit hole that actually taught me the problem. Starting April 15th.”
Specific date. Honest emotion — nervous, not thrilled. A real problem named plainly. A memory that doesn’t come with a polished lesson stapled to it. This reads like someone who actually exists and has a commute and ate lunch today.
Example 2: The Thought Leadership Post
BEFORE (AI Draft):
“The future of remote work is here. After delving into the latest research on distributed teams, here are three insights: 1) Trust is the new currency 2) Asynchronous communication saves time 3) Culture scales differently. Companies that understand these trends will thrive. Those that don’t will struggle. The choice is yours.”
AFTER (Human Version):
“Our Slack went down for six hours last Thursday. No channels. No DMs. No async anything. Everyone panicked — then something weird happened. We actually called each other. Talked on Zoom like it was 2012. Productivity went up that day. We cleared three things that had been stuck for two weeks. I’m not saying kill Slack. I’m saying if your remote team completely falls apart the moment it goes offline, you haven’t built a remote team. You’ve built a Slack team. Those aren’t the same thing.”
Same topic. Completely different approach. One starts with a specific Thursday. One ends with a real opinion — not a vague prompt to reflect or a soft call to action asking people to “share their thoughts below.”
The One Structural Move That Makes Posts Feel Personal
Start with a moment. Not a trend. Not a question. A moment.
A moment is something that happened to you — on a specific day, describable in concrete terms. A conversation. A number you saw on a dashboard. A mistake at 4:45pm on a Friday. A weird email from a client you still don’t fully understand. A time you were confidently, embarrassingly wrong.
The moment doesn’t need to be dramatic. It just has to be real. When you start there, your brain is recalling a memory — not constructing a concept. That difference shows up in your word choices. You’ll use specific language. You’ll keep irrelevant details your brain actually held onto — what you were drinking, who was sitting across from you, what the weather was doing outside the window.
Those irrelevant details? That’s what authenticity actually looks like. AI strips them out — it’s optimizing for clarity and structure, cutting everything that doesn’t serve the point. Humans keep them because they’re true. That’s what makes the moment endearing to us readers. We recognize it as something that genuinely happened to another person, not a composite lesson assembled by a language model at 11pm.
How to Use AI to Draft Without Sounding Like You Used AI
So, without further ado, let’s dive in — because the answer isn’t to stop using AI. It’s to stop using it as a finished product.
While you won’t need to throw away every AI draft entirely, you will need a handful of honest rewrites before anything is ready to post. First, you should generate the post and let it be as generic as it wants to be — at least if you want to get the structure out of your head quickly. That’s all the first draft is for.
Then delete the introduction. It’s always the worst part. Write your own opener starting with a specific moment or detail you actually remember. A number. A name. A date. Something pulled from actual memory, not assembled from training data.
Find the filler sentences next. “It’s important to note.” “In conclusion.” Any sentence that’s just quietly restating something you already said. Cut them — all of them. Then inject your actual opinion. Not a diplomatic take. Not a both-sides observation designed to offend no one. What do you actually think? Where do you disagree with the obvious answer? AI drafts toward consensus. You need to draft toward your specific, possibly controversial, actually-held position.
Adding something only you would know might be the best move of all, as LinkedIn content requires specificity that pure AI simply cannot manufacture. That is because it doesn’t have your project names, your actual revenue numbers, your colleague’s first name, the decision you made last March that surprised your whole team. Drop one of those in. It’s impossible to fake and it reads like nothing else on the platform.
Read it aloud before you post. If you stumble — or if it sounds like a memo from HR — rewrite it like you’re talking to someone you actually know. If you wouldn’t say it in a hallway conversation, don’t post it. The goal isn’t to trick anyone into thinking you don’t use tools. The goal is to actually sound like yourself. AI is scaffolding. You do the building.
Leave a Reply