How to write LinkedIn posts in your own voice with AI
To write LinkedIn posts in your own voice with AI, give the tool two separate things: a few examples of how you write, and the facts it may use for this post. Then edit the draft for accuracy, for detail and for phrases you would actually say. “Sound friendly” is not a voice. Your own sentences are.
Build a short voice brief
Pick three to five things you wrote and still like. Mix them up: a short post, a longer explanation, a reply to a client. If every example has the same shape, every draft will too. Strip out client names and anything private.
Then write a few lines about how you write:
- Audience: whose problem are you explaining?
- Words: which words do you use, and which would you never say?
- Rhythm: short and punchy, longer and patient, or a mix?
- Point of view: what do you believe about your work that others in your field don’t?
- Off limits: claims, client details and personal topics you won’t post about.
The free LinkedIn voice-brief builder walks you through these fields and turns them into a prompt you can paste into any AI assistant.
Keep style and facts apart
Your old posts contain old stories, old numbers and other clients. If you hand them over as “examples”, the AI will happily reuse those facts in a new post. So label them clearly: these examples are for style only. Then give it a short source note with the facts it may use this time.
Write a LinkedIn post for [specific buyer] about [one problem]. Use the writing examples only for style. Use the source notes as the only factual basis. Do not invent clients, results, quotes, numbers or events. If an important fact is missing, ask me or leave it out. Make one point and explain it with a concrete detail. Avoid the phrases listed in my voice brief. Return the draft and a short list of claims I should check. Voice brief: [paste] Writing examples: [paste] Source notes for this post: [paste]
Example: a weak draft and a grounded draft
Say you run a small web agency (example). Your source note reads: “When I review a client’s brief, I ask who signs off, what decision the page needs to support and what content already exists. Those questions come before any design talk.”
A weak draft: “We’ve transformed hundreds of brands with our three-step framework.” Your note never mentioned hundreds of brands or a framework. The AI made them up.
A grounded draft: “Before I talk about a website’s design, I ask three questions. Who signs it off? What should a visitor decide on this page? What content already exists? The answers show what the brief is missing, and the design conversation gets much easier.”
Every claim in the second draft traces back to your note. That makes it yours to post.
How to fix an AI LinkedIn post that sounds generic
“Make it sound more human” rarely works. Name the problem and give the fix.
The opening could belong to anyone
Swap the broad claim for the decision your reader faces. Instead of “Communication is the key to success”, try “Before the kickoff, agree who gives final approval.”
Tell the AI: “Start with the decision in my source note. Cut the introduction. Don’t add a story.”
It uses words you would never say
Keep a list of swaps from your own edits. If you would say “figure out what to do next”, ban “unlock actionable insights”.
Tell the AI: “Use the vocabulary in my examples. Replace the phrases I marked with everyday words.”
It sounds confident but says very little
Add the reason behind the advice. “Ask who signs off” is a start. “That person settles conflicting feedback, so the project stops going in circles” is what makes it stick.
Tell the AI: “Use this reasoning to explain why the advice matters: [your reasoning]. Keep it to one point.”
It adds experience you didn’t give it
Delete invented clients, quotes, numbers and first-person stories. A post doesn’t need a big result to be worth reading. A clear explanation of a real decision is enough.
Every post has the same shape
Change the job of the post, not just the opening line. Explain a mistake one week, walk through a method the next, help someone make a decision after that. The guide to repurposing LinkedIn content shows how one lesson becomes three different posts.
The ending turns into a sales pitch
End with a question that follows from the lesson, such as “Who settles conflicting feedback on your team?” Or simply stop when the point is made.
Review every draft in three passes
- Facts: can you trace every specific claim to something you supplied? Cut anything you can’t.
- Substance: will the reader make a decision or do something differently after reading?
- Voice: read it out loud. Rewrite any line you wouldn’t say to a client on a call.
Save your best edits as new examples. Over time your brief gets sharper and the first drafts need less work.
Keep your posts and your messages in the same voice
A post explains an idea to many people. A message answers one person. Keep the same words and the same expertise, but change the job. Pasting a polished post into a DM makes the conversation feel rehearsed.
Say your post explains why a project brief needs one decision-maker (example). A reader asks: “What if two partners both want final sign-off?” A good reply answers that question and nothing else:
“I’d ask them to agree which decisions each of them owns before the work starts. Is the disagreement about the direction, or about who makes the final call?”
Short, specific, and it sounds like the person who wrote the post. For more on that first exchange, see personalized LinkedIn message examples.
Where bravabase fits
bravabase does this routine with you. Brava, your personal AI sales agent, drafts posts in your voice from your ideas, notes and past posts. You edit with AI: ask it to rewrite or shorten a line, or describe the change you want in plain words. You can add an AI image or a carousel, then schedule the post on LinkedIn from your content calendar.
When someone replies, the conversation lands in your inbox with a drafted reply for you to review and send. See how content creation works. Drafts and chat are free; publishing a post uses credits (pricing).
Questions about writing LinkedIn posts with AI
How do I make an AI-written LinkedIn post sound like me?
Give it examples you wrote, a short description of your style and separate notes for the new post. Then correct the draft with specific instructions and read it aloud. Rewrite anything you wouldn’t say to a client.
What if I don’t have many published posts?
Use anything you’ve written yourself: an email, a proposal paragraph, notes from a talk. Or write a short answer to a question clients often ask, and use that as your first example.
How many writing examples should I give the AI?
Three to five is plenty. Choose pieces with different lengths and structures so the AI learns your voice, not one template.
Does the free voice-brief builder write the post for me?
No. It builds a prompt from the details you enter. You paste that prompt into the AI assistant you already use to get a draft.