Most AI writes LinkedIn posts that sound like AI.
You've tried ChatGPT. You've tried other AI writing tools. The posts come back polished, professional, and completely generic. Say Something is built differently — it has a kill list, an interview process, and rules about what it will never write.
You can spot an AI LinkedIn post from a mile away.
They all start the same way. A bold claim. A single-line paragraph for drama. An emoji or two. Then a listicle of “key takeaways” followed by “repost if you agree.” The person posting it could be anyone, because the content says nothing specific about them.
The issue isn’t that AI wrote it. The issue is that nobody told the AI what not to write. So it defaults to the most common patterns on the platform — which happen to be the most annoying ones.
Thirteen things Say Something will never put in your post.
Banned from every draft
- Emojis anywhere in the post
- Hashtag stacking at the bottom
- “I’m humbled / honored / thrilled”
- “Repost if you agree”
- Em dashes for dramatic effect
- Single-line dramatic paragraphs
Also banned
- Rhetorical questions as hooks
- Listicles and numbered frameworks
- Engagement bait CTAs
- Clichéd motivational language
- Generic advice with no specifics
- Corporate jargon and buzzwords
What’s left? Posts that open with what happened, include real details, give credit where it’s due, and end when they’re done — not with a call to action. See the full breakdown.
Interview first. Write second.
Start with something real
A win, a lesson, a trend. You don't need a polished idea. Just something that happened to you, not a generic topic.
Short conversation
Two or three questions to pull out the details that make it yours. The specific names, numbers, and moments that no one else has.
Three human-sounding drafts
Short, medium, and long. Each one filtered through the kill list. Each one sounds like a person who was actually there.