Claude Prompts for LinkedIn Posts That Don't Sound Like AI

The best Claude prompts for LinkedIn posts give Claude your raw facts, a short voice brief and a clear post type, then ask for a plain first draft you edit yourself. Feed it specifics, ban the usual AI phrases up front, and finish with an edit pass. The result reads like you on a good day, not a template.

Below are seven copy-paste prompts: a one-time voice setup, four post types with before and after examples, an edit pass that strips AI tells, and an ideas prompt. For the rest of your workday, see our 50 copy-paste prompts for work.

Why do AI-written LinkedIn posts sound like AI?

Most AI posts fail because the writer asked for "a LinkedIn post about my promotion" and nothing else. With no facts or voice to work from, the model falls back on the most common LinkedIn shape, and readers scroll straight past it. Watch for these tells:

Read every draft out loud. If it sounds like a press release, it is one.

1. The voice brief prompt (run this once)

What it is: a short profile of how you write, built from things you've already written. Why it matters: Claude can't sound like you if it has never seen your writing. Who it suits: anyone who plans to post more than once or twice.

I want help writing LinkedIn posts that sound like me, not like AI.
Below are 3 to 5 things I've written (posts, emails or chat messages).
Write a voice brief of 8 to 10 bullet points covering sentence
length, vocabulary, humor, how I open and close, and words I use
or avoid. Match my spelling (US or UK). Then list 10 phrases that
don't fit my style.

My writing samples:

If you use Claude's Projects feature, put the brief in the project instructions so every chat starts with your voice. Pair it with these prompts to optimize your LinkedIn profile so your profile and posts sound like one person.

2. The job update prompt

What it is: a prompt for announcing a new job, a promotion or a move on. Why it matters: job updates attract "new chapter" cliches, yet your network actually wants to read them. Who it suits: anyone with career news.

Using my voice brief, write a LinkedIn post announcing a job change.
Use only these facts; add nothing.
- What changed:
- What I'll be working on:
- The one thing I'm most looking forward to:
- One person or team I want to thank, and the specific reason:
Rules: under 150 words. Put the news in the first sentence.
No "thrilled", "humbled", "excited to announce" or "new chapter".
No emoji. No hashtags. Do not end with a question.

Before and after

Before (typical AI draft):

I'm thrilled and humbled to announce that I'm starting a new chapter as Senior Product Manager! This journey taught me so much about growth and resilience. Grateful for everyone who believed in me. Onward and upward!

After:

After four years running support operations, I've moved into product. Starting Monday I'm a Senior Product Manager on the billing team, working on the same tools my old team filed tickets about for years. Big thanks to Priya, who let me sit in on her roadmap reviews long before any of this was official.

Looking for work instead? Swap in the roles you want and the problems you solve, then see how to use AI in your job search.

3. The lesson learned prompt

What it is: a prompt for a post about a mistake, a surprise or something you'd now do differently. Why it matters: honest lessons start real conversations, but AI loves inventing a tidy moral. Who it suits: specialists, managers and freelancers who want to share expertise without bragging.

Using my voice brief, write a LinkedIn post about a lesson I learned.
What happened, in my own words:

What I'd do differently now:

Rules:
- Start in the middle of the moment, with one concrete detail.
- Use only the events I described. Do not invent dialogue,
  numbers or outcomes.
- State the lesson once, in one plain sentence, near the end.
- No "Here's what I learned", no "Let that sink in", and no
  one-line paragraphs stacked for drama.
- 120 to 200 words.

Before and after

Before:

I made a huge mistake last week. Let that sink in. I sent a report to a client without double-checking it. Here's what it taught me about leadership: ownership matters, communication is key, and mistakes are opportunities.

After:

Last Tuesday I sent a client a forecast with one formula pointing at the wrong tab, so every total below it was off. They caught it before I did. I called, walked them through the fix and resent it within the hour. I changed one habit: anything with numbers in it now gets a second pair of eyes before it leaves my inbox.

The "do not invent" line matters most. Claude fills gaps with plausible detail, which is exactly what makes a story feel generic. Anthropic's prompt engineering guide for Claude recommends clear, specific instructions, and that includes saying what you don't want.

4. The project showcase prompt

What it is: a prompt for posting finished work, such as a launch, a case study or a portfolio piece. Why it matters: real work is credible, but AI drafts turn it into an ad. Who it suits: freelancers, consultants and anyone whose work is visible.

Using my voice brief, write a LinkedIn post about a project I finished.
Facts:
- The project:
- Who it was for (name the client only if I say I have permission):
- The problem before:
- What I actually did, in 2 or 3 specific steps:
- The result, using only numbers I give you:
- One thing that didn't go to plan:
Rules: lead with the problem, not with me. No "game-changer", "seamless" or "excited to
share". Under 180 words. Give me 3 different first lines to pick from.

Before and after

Before:

Excited to share a game-changing project! We delivered a seamless, innovative website redesign that transformed our client's online presence. So proud of this amazing team! #webdesign #innovation #growth #success

After:

A local bakery's website loaded so slowly on phones that people gave up before the menu appeared. Instead of a redesign, we compressed every photo, removed plugins nobody used and turned the menu into a plain text page. It now opens in a couple of seconds on an ordinary phone. What didn't go to plan: the ordering tool had to stay as it was, because swapping it meant retraining staff right before the holidays.

Check your contract before posting client work, since many agreements limit what you can share publicly.

5. The hiring post prompt

What it is: a prompt for announcing an open role on your own team. Why it matters: a post from the hiring manager reads warmer than a job ad, but AI versions slide into "rockstar" territory. Who it suits: hiring managers, founders and team leads.

Using my voice brief, write a LinkedIn post announcing a role on my team.
Facts:
- Job title and team:
- Location and setup (remote, hybrid or on-site, and which countries):
- Pay range:
- What this person will do in their first 90 days:
- Two must-have skills and one nice-to-have:
- The honest reason someone would want this job:
- How to apply:
Rules: plain words a candidate would use. No "rockstar", "ninja",
"fast-paced", "wear many hats" or "we're a family". Avoid wording
that hints at age, gender or other personal traits. Under 170 words.
End with how to apply.

Before and after

Before:

We're hiring! Looking for a rockstar developer to join our fast-paced, family-like team. Must be a digital native who can wear many hats. DM me!

After:

I'm hiring a backend developer for our payments team, fully remote within the US or UK, with the pay range in the job post linked below. In your first three months you'd take over our refund service. You need solid Python and experience with a relational database; payments experience is a bonus, not a requirement. The honest pitch: a small team, documented code and no weekend on-call. Apply through the link, not my DMs.

Check the rules where you hire. Several US states, including California, Colorado and New York, require pay ranges in job postings, and both the US and the UK ban hiring discrimination based on protected characteristics such as age, sex and race. Cut phrases like "digital native", which can read as age bias, and check with HR if unsure.

6. The edit pass prompt that strips AI tells

What it is: a second prompt you run on any draft, including drafts you wrote yourself. Why it matters: it catches the stock phrases the post prompts miss. Who it suits: everyone, on every post.

Edit the LinkedIn post below so it sounds like a person wrote it.
Work in this order:
1. List every phrase that sounds generic, salesy or AI-written.
2. Remove long dashes, emoji, hashtag lists and "Agree?" endings.
3. Replace each abstract claim with the concrete fact from my
   notes, or cut it.
4. Merge one-line paragraphs into paragraphs of 2 to 3 sentences.
5. Cut the length by 20 percent without losing a fact.
6. Show the final version, then list any claim you could not
   confirm from my notes.
Do not add anything new.

The post:

Step 6 is your safety net: confirm or delete every flagged claim. LinkedIn's Professional Community Policies ask members not to share false or misleading content, whoever wrote the sentence.

7. The post ideas prompt for when you're stuck

What it is: a way to get LinkedIn post ideas with AI that come from your real work, not from trending topics. Why it matters: the hardest part of posting is usually deciding what to say. Who it suits: anyone who goes blank at the empty box.

Interview me to find LinkedIn post ideas. Ask me 8 questions, one
at a time, about my last month at work: problems I solved, mistakes,
questions people asked me, tools I changed and opinions I disagree
with. After my answers, give me 10 post ideas. For each, name the
post type (job update, lesson learned, project showcase, hiring or
opinion) and suggest a first line in my voice. Only use topics I
mentioned in my answers.

Because the ideas come from your answers, they're specific by default. For more AI prompts for LinkedIn posts and other work tasks, browse the Boredlisted prompt library.