LinkedIn Post Prompts for ChatGPT That Don't Read Like Everyone Else's Feed

The short version: Good LinkedIn post prompts for ChatGPT give it a specific structure, a real detail from your week, and a clear point of view, not just a topic. The prompt gets you a usable skeleton in thirty seconds, but the sentence you post is the one you rewrote after reading the draft out loud. I'll give you the exact prompts I use, why the obvious ones flop, and the bit about ChatGPT and LinkedIn that most prompt-pack sellers won't tell you.

Why "write me a LinkedIn post about X" is a waste of a good idea

I've watched this happen to clients so many times it's boring to type out. Someone has a interesting week, a client win, a mistake, a strong opinion about their industry, and they type "write a LinkedIn post about how AI is changing marketing" into ChatGPT. Out comes 200 words that could have been written by anyone, about anything, for anyone. Three emoji bullet points. A line that says "the future is here." A call to action asking people to "share your thoughts below."

That post gets 4 likes. Two of them are from people who like everything you post out of loyalty.

The problem isn't ChatGPT. It's the prompt. You gave it a topic instead of a shape, and it filled the gap with the most statistically average LinkedIn post it's ever seen, because that's what it's trained to do when you leave it open-ended. If you want something that doesn't sound like it was written by the same tool as the other 40 posts in someone's feed that morning, you have to be more specific than the topic. You have to give it a job.

The prompt structure that gets you a usable draft

Every prompt I use has four parts, in this order:

• The format (story, opinion, list, contrarian take, mistake post)
• One real detail only I would know
• The exact length and tone
• What I want the reader to feel or do at the end

Here's the difference in practice. Weak prompt: "Write a LinkedIn post about client retention." Working prompt: "Write a LinkedIn post in the story format. Open with the exact moment I realised a client of 6 years was about to leave because I hadn't sent an invoice update in 4 months, not because of the work. Keep it under 150 words, first person, no corporate language, end with one direct question for the reader, not a generic call to action." The second one gives ChatGPT almost nothing to invent, which is exactly why it works. It can only draft around the facts you fed it.

Prompt 1: the story post

"Write a LinkedIn post about [specific event, with the date or rough timeframe and one concrete number]. Structure it as: one-line hook that states the outcome first, then the story in past tense, then one lesson stated plainly, no more than 4 short paragraphs, no hashtags, no emoji, British English."

Example I used in November 2025: "Write a LinkedIn post about the time I turned down a £40,000 retainer in 2019 because the client wanted me to ghostwrite content I didn't believe in. Hook first with the number. Then the story. Then the lesson: fees you regret taking cost more than the ones you turn down. Under 180 words." That post got 89 comments, more than any post I'd put up in the previous six weeks, because the hook led with a real number and a real refusal, not a platitude.

Prompt 2: the opinion post

"Write a LinkedIn opinion post arguing that [your specific, slightly uncomfortable position]. State the opinion in the first sentence, no throat-clearing. Give one piece of evidence from my own work. Address the obvious counterargument in one sentence. End flat, no question, no hedge."

Use this when you disagree with the industry consensus, not when you're trying to manufacture disagreement for engagement. If you don't believe the opinion, it shows in the third paragraph, and readers on LinkedIn are better at spotting fake conviction than most people give them credit for.

Prompt 3: the mistake or lesson post

"Write a LinkedIn post admitting a specific mistake I made [in this context], the real consequence it had, and what I do differently now. No self-flagellation, no 'this taught me humility' cliches, just the facts and the fix. Around 130 words."

These consistently outperform polished win posts on my feed. My most-shared post of 2025 wasn't a client result, it was 90 words about losing a £6,000-a-month retainer because I stopped checking in monthly once the work was "working." That one got picked up and reposted by three people whose accounts I didn't even follow.

Prompt 4: the list or framework post

"Write a LinkedIn post structured as a numbered list of [3 to 5] things about [topic], each one no more than two sentences, each one containing a specific number or example, not a general principle. Title line as the hook. No summary paragraph at the end."

This is the closest ChatGPT gets to writing a tight post without help, because the list structure forces short sentences. It still needs you to swap in real numbers, because ChatGPT's default instinct is to write "many businesses see improved results" instead of "clients using this saw response rates go from 2% to 11% over 8 weeks."

Prompt 5: the contrarian take

"Write a LinkedIn post that challenges a common belief in [my industry]: that [the common belief]. Open by stating the common belief, then say plainly why it's wrong or incomplete, using one specific example from my experience. Keep the tone direct, not smug. Under 150 words."

Use sparingly. If every post is a contrarian take, you stop being interesting and start being predictable, which is its own kind of boring.

A real example from my own feed

In September I posted about turning down a speaking gig that would have paid £3,000, because the organiser wanted a "safe, inspirational" talk with no mention of the five years my business shrank before I rebuilt it. I ran the story through ChatGPT with the story-post prompt above, got a draft in about 20 seconds, and then spent nearly 25 minutes editing it. The original draft had three phrases I never say out loud: "on my journey," "at the end of the day," and "I truly believe." I cut all three. I changed the ending from a question ("What would you have done?") to a flat statement ("I'd rather speak to 40 people who need the real version than 400 who want the safe one"). That edited version got 156 reactions and 31 comments, more than double what my average post gets. The prompt got me the bones. The 25 minutes of cutting got me the post that sounded like me.

The uncomfortable bit nobody selling a prompt pack wants to say out loud

Here's what I've noticed after two years of using ChatGPT on LinkedIn content most days: the prompt is maybe 20 percent of what makes a post good. The other 80 percent is you sitting there afterward, reading it out loud, and cutting every sentence that sounds like it was written to please an algorithm instead of a person. Anyone telling you a perfect prompt will get you a post you can copy and paste straight onto LinkedIn is either selling you something or hasn't posted enough to know better.

I'd also say something people rarely admit: LinkedIn's own audience has got faster at spotting the tell-tale rhythm of AI-written posts, the three-word punchy sentence followed by a longer one, the "here's the thing" transitions, the tidy three-item lists that appear even when the topic doesn't naturally break into three. If your prompt produces something that reads smoothly but says nothing you couldn't have guessed, that smoothness is now working against you, not for you. The posts doing well on my feed right now are rougher round the edges, with a typo left in occasionally, a sentence that trails off, a number that's oddly specific. That's not an accident. If you want prompts specifically built to avoid that AI cadence, I've gone deeper on exactly that problem in this piece on ChatGPT prompts for LinkedIn posts that don't sound like AI, with more examples of the phrases to strip out.

How I use these day to day

My process, most weeks, looks like this:

• Monday morning I jot down 3 to 5 real things from the past week: a client conversation, a number that surprised me, a mistake, a small win, something I disagreed with in a webinar
• I pick the format that fits each one (story, list, opinion, mistake) rather than forcing every idea into the same shape
• I run one prompt per idea, using the structure above, and get a first draft in under a minute
• I read each draft aloud, cut anything I wouldn't say to a client over coffee, and cut every sentence that starts with "In today's fast-paced world" or similar
• I schedule the posts across the week rather than posting all five ideas on the same day, because variety in timing matters as much as variety in format

That's roughly 15 minutes of prompting and 20 to 30 minutes of editing per post, four or five posts a week. It's not fast in the way "AI writes your content for you" adverts promise, but it's a lot faster than staring at a blank post box for 40 minutes, which is what I was doing before.

Prompts for specific situations

If you're job hunting and want to post about your search without sounding desperate: "Write a LinkedIn post announcing I'm open to work as [role], mentioning one specific project I'm proud of with a number attached, no generic 'passionate about' language, ending with what kind of company I'm looking for, not a plea for help." This pairs well with the interview prep side of things too, if you're using ChatGPT to get ready for conversations that might come from the post, and I've broken that down separately in this guide to ChatGPT prompt ideas for job interview practice.

If you're promoting a piece of content you made elsewhere, like a YouTube video, don't just paste the title and ask for a caption. Give ChatGPT the actual hook from the video and one specific viewer reaction if you have one; I use a similar detail-first approach when I'm scripting the video itself, which I cover in this piece on YouTube script prompts for ChatGPT.

If your job is specific enough that generic marketing prompts don't fit (you're in law, healthcare admin, finance, teaching), the format above still works, but the detail you feed it needs to be your industry's specific detail, not a marketing one. There's a broader set of role-specific starting points in this list of ChatGPT prompts by job across 12 professions if you want to see how the same structure adapts.

And if you're cross-posting the same story to Instagram, don't reuse the LinkedIn draft word for word. The tone that works on LinkedIn reads stiff on Instagram, and the caption prompts need a different rhythm entirely, which I go into in this piece on Instagram caption prompts for ChatGPT.

What to stop doing

Stop asking ChatGPT to "make it more engaging." That instruction almost always produces more exclamation marks, more rhetorical questions, and more hollow enthusiasm, none of which is what engagement runs on. Stop asking for hashtags at the end; LinkedIn's own data has shown hashtags make very little difference to reach now, and three random ones stapled to the bottom of a post look like an afterthought because they are one. Stop letting it write your hook last; the first line is the only line most people read before deciding whether to open the rest, so write that yourself, by hand, after the draft exists, using the sharpest true fact you have.

If you want a rotating supply of fresh prompts to test rather than reusing the same five forever, I update a running list of newer ones worth trying each month in this collection of new ChatGPT prompts, because what performs on LinkedIn shifts as more people catch on to whatever's working right now.

A closely related walkthrough: AI Prompts for Marketing That Work (Not the Copy-Paste Lists Everyone Shares).

This guide is part of my ChatGPT Prompts and Limits: 30 Guides to Fun Prompts, Photo Edits and Daily Caps. To write your own, try the free ChatGPT prompt generator.

Frequently asked questions

Can ChatGPT write a LinkedIn post that sounds like me instead of generic?

Not on its own. It can get close if you feed it a real detail, a specific number, and a clear format instruction, but the sentences that make a post sound like you rather than a template still need a human edit, usually 15 to 25 minutes of cutting corporate phrases and softening the AI cadence.

How long should a LinkedIn post prompt for ChatGPT be?

Longer than most people write. A good prompt runs 3 to 4 sentences covering format, one real detail, length and tone, and the intended reader reaction. A one-line topic prompt like "write about leadership" gives ChatGPT nothing to work with except averages.

Should I tell ChatGPT to write in my voice?

Telling it isn't enough; showing it works better. Paste in two or three of your best-performing past posts as examples before asking for a new one, and it will match your sentence length and rhythm far more closely than if you just describe your tone in words.

Why do my ChatGPT-written LinkedIn posts get less engagement than my old ones?

Usually because they're too smooth. LinkedIn audiences have got quicker at recognising the tidy three-part structure and safe language AI defaults to, and posts that read a bit rougher, with a real number, a specific mistake, or an unfinished thought, are outperforming the polished ones in most feeds right now.

Useful references

• LinkedIn's Help Centre
• OpenAI's own platform documentation