Humanize your LinkedIn post before it posts you.

The LinkedIn-specific tells

Beyond the general blacklist, LinkedIn AI drafts share a uniform: an emoji-led hook (🚀✨💡), 'excited to share', a numbered list of platitudes that could end in any order, bolded buzzwords, an engagement-bait question, and hashtag confetti. Readers scroll past this shape on sight now — the format itself is the tell, before a single word is read.

The rewrite recipe

1. Replace the topic with a story. 'Thoughts on leadership' is a topic; 'I promoted the wrong person' is a post. 2. One idea only. Five lessons teach nothing; one specific mistake teaches plenty. 3. First line = the hook, no emoji. You have ~140 characters before 'see more' — spend them on the most concrete sentence you have. 4. End with a statement, not a survey. 'What's YOUR journey?' begs; a strong closing line earns comments. 5. Run the checker; anything over ~5 tells per 1,000 words needs another pass — see the gallery example for the full transformation.

Hook surgery: first lines, before and after

The pattern: every humanized hook contains at least one number, one concrete noun, or one admission — and zero emojis.

The shapes that still work (and the one that's dead)

Four post shapes keep earning reach in 2026 precisely because they're hard to fake: the specific failure (one mistake, its cost, the fix — our gallery example), the small dataset ("we A/B-tested 200 cold emails; here are the three subject lines that won"), the contrarian-with-receipts (a common belief, your evidence against it, the nuance), and the process teardown (exactly how you do a thing, with real numbers, no gatekeeping). The dead shape is the one AI produces by default: the virtue listicle — five generic lessons, an engagement question, hashtag confetti. It still gets impressions from the algorithm; it no longer gets respect from the humans who could hire you, which was the point of posting.

Comments are the second post

One more tell nobody audits: AI-drafted replies under a human post. "Great insights! Thanks for sharing 🙌" under your own thoughtful post undoes it. Reply like you write — reference the specific point, add one thing, disagree where you disagree. The algorithm weighs comment threads heavily, and readers judge the author by both halves of the conversation.

A worked example, start to finish

The AI draft: "🎯 5 Tips for Better Time Management! In today's busy world, managing your time effectively is more important than ever. 1) Prioritize your tasks 2) Eliminate distractions 3) Take regular breaks 4) Use productivity tools 5) Learn to say no. Which tip resonates with YOU? Share below! #productivity #timemanagement #growth"

The 4-minute humanize: pick the one tip you have actual evidence for (say, #5), find the story behind it, cut the rest. Result: "I said yes to every meeting for a year — my calendar says 611 of them. Then I tried a rule my old boss used: every yes must name what it displaces. 'Yes to this call' means 'no to the proposal draft.' Saying the trade-off out loud killed a third of my meetings in a month, and nobody complained. Turns out most invitations are habits, not needs." Same topic, one idea, one number, one borrowed-and-credited trick — and nothing for the feed's pattern detector to yawn at.

Format mechanics that quietly matter

Line breaks: AI drafts default to dense paragraphs; the feed rewards 1–2 sentence blocks with air between them — but breaking every sentence into its own line is now its own tell (the "broetry" uniform). Mix block lengths like you mix sentence lengths. Hashtags: three specific ones outperform confetti; none at all is increasingly common among strong writers. Links: the algorithm demotes external links in the post body — putting the link in the first comment is standard practice, and your post text should work without it. Length: 900–1,200 characters is the engagement sweet spot for story posts; AI drafts either stop at 400 (nothing said) or sprawl past 2,000 (everything padded). Editing window: reread 10 minutes after posting — the tells you missed at draft time glow in the published version, and early edits don't hurt distribution.