ChatGPT Prompts for Writing LinkedIn Posts That Go Viral

The short version: ChatGPT can write you a LinkedIn post that gets a lot of views, but "viral" is a distribution outcome, not a writing style, and the prompts that work are the ones that force ChatGPT to use your real experience, your real numbers, and your real opinions instead of generic advice. I've tested dozens of these prompts on my own account since 2023. The ones below are the ones that produced actual results, not the ones that sound clever in a listicle.

Why "write me a viral LinkedIn post" doesn't work

I typed exactly that into ChatGPT once, in 2023, as an experiment. What came back was a post about "the power of failure" with three bullet points and a line that said "what's your take?" I posted it anyway, just to see. It got 11 likes. My cat has more Instagram followers than that post got engagement.

The problem isn't ChatGPT. The problem is the prompt. "Write a viral post" gives the model nothing to work with except the internet's average idea of what a viral post looks like, which by 2026 is a well-worn template: short lines, a bold claim, a listicle, a question at the end. LinkedIn's algorithm has seen that template a hundred million times. It doesn't reward familiar shapes any more, it rewards specific stories and strong opinions that make people stop scrolling.

The prompts that get results give ChatGPT three things: a real detail from your week, a number, and a stance you're willing to defend in the comments. Take those three away and you get filler, no matter how good the prompt sounds.

The prompts I use (copy these)

These aren't polished for a blog post. This is what I paste into ChatGPT, more or less word for word.

1. The "uncomfortable truth" prompt

"I work in [your field]. Ask me five questions to find an opinion I hold about [your field] that most people in it would disagree with. Then write a 150 word LinkedIn post arguing that opinion in my voice, first person, one idea per line, no hashtags, no emojis, ending with a direct question."

This works because it forces ChatGPT to interview you before it writes anything, which means the post is built on your actual belief rather than a generic template. If you want the finished draft to sound like you rather than like every other LinkedIn post from the last two years, this is closer to the approach I break down in these prompts for LinkedIn posts that don't sound like AI.

2. The "specific number" prompt

"Here's a result I got this week: [paste the actual number, e.g. 'went from 40 to 340,000 impressions on one post']. Write a LinkedIn post that opens with that exact number in the first line, explains in plain terms what I did differently, and includes one thing that didn't work before it worked. Keep it under 200 words."

Numbers stop the scroll. "I grew my audience" gets skipped. "I went from 200 followers to 14,000 in four months" makes people click "see more." Always feed ChatGPT the real number. If you don't have an impressive one yet, use a small honest one, "I sent 30 cold emails and got 2 replies" beats a made-up big claim every time, because readers can smell invention from three sentences away.

3. The "disagree with an expert" prompt

"Find a piece of common advice given in [your industry] that you think is outdated or wrong. Write a LinkedIn post where I politely but firmly disagree with it, give one real example from my own work that contradicts the advice, and invite people to push back in the comments."

This is the prompt behind most of the posts on my feed that pull 200-plus comments. Agreement gets a like. Disagreement gets a comment, and comments are worth roughly ten times more to LinkedIn's algorithm than likes in the first hour, based on the pattern I've watched across my own posts for the past three years.

4. The "before and after story" prompt

"Turn this into a LinkedIn post: [paste a rough, messy paragraph describing something that changed for you, a client, or your business]. Structure it as: the moment before it went wrong or right, the specific thing that happened, the specific number or result, and what I'd tell someone in the same position now. Write it in short paragraphs, first person, no corporate language."

Feed it something real and unpolished. The messier your input, the more human the output, because ChatGPT has actual texture to work with instead of having to invent a story from scratch.

5. The "steal my framework" prompt

"I have a 3-step process I use for [task]. Here are the steps: [list them roughly]. Write a LinkedIn post that teaches this process in under 150 words, uses a numbered list, and ends by telling people they can copy it exactly, no need to credit me."

Generosity performs. Posts that give away a usable framework, one people can screenshot and use tomorrow, get saved and shared far more than posts that tease a lesson and pitch a course at the end.

A real example from my own feed

In January 2026 I posted a piece about a client call that went badly. I'd pitched an AI workshop to a manufacturing firm in the Midlands and the finance director asked me, live on the call, "what's the actual return on this, in pounds, by month three." I didn't have a clean answer ready. I froze for about four seconds. That's the whole story.

I ran that story through the "before and after" prompt above, added the real detail that I lost the contract, and posted it with the headline line "I lost a £14,000 contract because I couldn't answer one question." That post got 1,200 reactions and 96 comments, which for my account at the time was roughly six times my average. Compare that to a post I put up the same week using the generic "5 tips for AI adoption" format, built from a template prompt, which got 43 reactions and two comments.

Same account, same week, same audience. The only difference was one post admitted I got something wrong and gave a specific number, and the other one gave advice nobody asked for. That's the whole lesson, really. ChatGPT didn't make the first post work. My willingness to include the embarrassing detail made it work, and the prompt just got that detail onto the page faster than I could have typed it myself.

The uncomfortable part nobody selling prompt packs wants to say

Here's the bit most "50 viral ChatGPT prompts" posts skip over. No prompt makes a boring account go viral. If your last ten posts got under 100 views each, feeding ChatGPT a better prompt will get you a better written post, not a differently performing one. Distribution problems and writing problems are different problems, and prompts only fix the second one.

I say this as someone who spent five years rebuilding an audience from close to nothing after stepping back from an influencer career that had already peaked once. The prompts helped me write faster. They didn't fix the fact that for the first three months back, almost nobody saw what I posted, because LinkedIn was still deciding whether my account was worth showing to anyone. If a handful of posts in a row underperform, the issue usually isn't the prompt, it's something structural with the account or the posting pattern, and I've written a full diagnostic on that over on why your LinkedIn post is getting no views and how to fix it, which is worth reading before you blame the writing.

The other uncomfortable truth: chasing virality directly usually backfires. The posts that spread wide, in my experience and everyone else's I've compared notes with, were rarely the ones written to go viral. They were posts written to make one specific point clearly, using a real story, that happened to resonate with more people than expected. Optimise for clarity and honesty first. Treat any spread beyond that as a bonus, not the goal, because writing for the algorithm instead of the reader is the fastest way to produce something that reads exactly like everyone else's AI output.

Step by step: from blank page to posted, in fifteen minutes

1. Write one messy sentence about something that happened to you this week, work related, no polishing. Example: "client asked me a question I couldn't answer and I lost the deal."
2. Paste it into ChatGPT with this instruction: "Ask me four follow-up questions to get the full story before writing anything."
3. Answer honestly, including the number if there is one, money, time, followers, whatever it is.
4. Ask it to draft the post using one of the five prompts above, specifying under 200 words, first person, no emojis, no hashtags.
5. Read the draft aloud. Cut any sentence you wouldn't say to a friend over coffee.
6. Change the opening line so it starts with the number, the mistake, or the disagreement, not a scene-setter like "last week I had an interesting conversation."
7. Post it, then reply to every single comment within the first hour. This step matters more than the prompt did.

That last step is the one people skip and it's costing them reach. LinkedIn's early engagement window is short, roughly the first 60 to 90 minutes, and replying to comments during that window signals to the algorithm that the post is generating a real conversation, which is what gets it pushed to a second, wider batch of feeds.

Adjusting prompts for your specific role

Generic prompts produce generic posts even with a good structure, so the second layer is telling ChatGPT exactly who you are and what you do before you ask for anything. A recruiter's viral post looks nothing like a consultant's viral post. If hiring is your world, the same "specific number, real story" logic applies to your job adverts too, and I've laid out the exact prompts for that over in these ChatGPT prompts for recruiters writing job adverts that get applications. If you're building a personal brand from scratch rather than promoting a company, there's a longer list of angles worth working through in these 30 ChatGPT prompts for personal branding posts, and if your strongest content is client results rather than opinions, the case study format tends to outperform the hot take format, which is covered in these ChatGPT prompts for writing case studies.

Whichever direction you go, tell ChatGPT your job title, your audience, and one recent, specific thing that happened, every single time you prompt it. Skip that step and you're back to the "power of failure" post with 11 likes.

If you'd like to see how the options stack up, read Taplio alternatives.

What "going viral" costs you afterwards

One thing nobody mentions: a post that does well, say 300,000-plus impressions, brings a wave of connection requests, DMs pitching you services you don't want, and comments from accounts trying to piggyback on your reach. My best post of 2026 so far pulled in around 40 new connection requests a day for a week, and maybe three of them turned into anything useful. Virality is not the same as leads. Plan for the noise before you post, not after.

I keep every related walkthrough in the ChatGPT Prompts and Limits: 30 Guides to Fun Prompts, Photo Edits and Daily Caps. To write your own, try the free ChatGPT prompt generator.

Related: chatgpt (this site).

Related: manufacturing: guidelines and how to pitch.

I've also covered a LinkedIn profile that attracts clients. I've also explained LinkedIn profile tips that attract clients.

Frequently asked questions

Can ChatGPT write a LinkedIn post that goes viral on its own?

No. ChatGPT can write a well-structured, clear post fast, but virality depends on the account's existing reach, the strength of the real story or opinion behind the post, and how quickly you engage with comments after posting. Treat ChatGPT as a drafting tool, not a growth strategy.

What's the single biggest mistake people make with viral LinkedIn prompts?

Asking ChatGPT to invent a story or opinion instead of feeding it a real one. Posts built on invented anecdotes read flat because there's no specific detail underneath them. Always give the model a genuine, even embarrassing, detail to work from.

How long should a LinkedIn post be to get good reach?

Shorter usually wins for opinion and story posts, aim for 150 to 200 words. Longer posts, 400 words plus, can work for teaching content or case studies but need strong line breaks and a reason to keep reading past the first three lines, since that's roughly where LinkedIn cuts the preview.

Should I post the ChatGPT draft exactly as written?

No. Read it aloud and cut or rewrite anything that doesn't sound like something you'd say out loud. The drafts that perform best are ones where you've changed at least a third of the wording after ChatGPT's first pass, usually the opening line and the ending question.

Primary sources

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