A LinkedIn ghostwriter writes posts that go out under your name: they collect material from you, usually in short interviews, and turn it into posts that sound like you. AI can take over the writing part if it gets a voice profile, real examples of your writing and your raw material. What it can't supply is the experience itself, the follow-up questions that draw it out of you, and the responsibility. The same LinkedIn rules apply either way: you are responsible for every post, you don't share your account with anyone, bots may not post, comment or like for you, and LinkedIn recommends saying so when you relied heavily on AI, unless that is obvious from the context.
My LinkedIn posts are drafted by an AI employee, Conny: short posts on AI news, each with an image that shows me as a recurring pixel-art figure. She writes in my voice, because I post as myself, not about myself in the third person. I approve every post. Everything on the topic is collected on my LinkedIn with AI page. This article answers one question: a person, an agency or AI?
What a LinkedIn ghostwriter actually does
Ghostwriting means someone else writes and you are the sender. Depending on the package, it covers:
- Collecting material: calls or voice notes where you talk about projects, mistakes and opinions.
- Planning topics: what you write about, in which order, for which readers.
- Writing: drafts in your voice, often with several openings to choose from.
- Formats and review: text or document posts, and which ones get read.
The real value is rarely the wording, it's the first item: a human asks again when you say "it went well" and notices the story hiding in a side remark.
What LinkedIn ghostwriting costs
Our research turned up no neutral, public price list. Some providers quote only after a first call, and most ranges online come from sellers of ghostwriting or writing tools. That's why this article gives no number (as of November 2026). Ask every provider the same questions:
The LinkedIn rules that apply to ghostwriters and AI
Who writes doesn't change the rules. Four matter, all as of November 2026.
Your account stays yours. The User Agreement says: "You will not share your account with anyone else". The Professional Community Policies add: "do not share your LinkedIn account with anyone else." A ghostwriter who logs in with your password and posts for you doesn't fit those rules. The clean setup: they deliver the draft, you publish it. LinkedIn lets you schedule a post between 10 minutes and 3 months ahead, so approving and scheduling become one step.
No bots for posts and reactions. The User Agreement prohibits using "bots or other unauthorized automated methods" to, among other things, "create, comment on, like, share, or re-share posts, or otherwise drive inauthentic engagement". AI can suggest a reply to a comment. You send it yourself.
No manufactured engagement. The Professional Community Policies say: "Don't do things to artificially increase engagement with your content." They also ask you not to agree with others ahead of time to like or re-share each other's content. An "engagement group" add-on has no place in a ghostwriting package.
Your experience, your responsibility. LinkedIn's guidance on AI content says content "should still reflect the member's own voice, perspective, and experience" and "You are ultimately responsible for everything you post on LinkedIn". It recommends a note if you relied heavily on AI, unless that's obvious from the context. And the Community Policies rule out misleading information about your qualifications or achievements: no ghostwriter, human or AI, should invent a win for you. Whether the law requires disclosure beyond LinkedIn's recommendation is covered in AI Disclosure Rules Under the EU AI Act. That's my perspective as an operator, not legal advice.
LinkedIn also says its focus "is not on how content is created, but whether it adds value": posts that feel generic or lack a clear point of view are less likely to be widely distributed, and members can flag them as "Seems like AI slop". How LinkedIn distributes posts is explained in How the LinkedIn Algorithm Works in 2026.
When AI can replace a ghostwriter, and when it can't
The line isn't between good and bad writing. It's about the material. An assessment, not a measurement:
AI is enough if you supply raw material regularly: voice notes, bullet points, call transcripts. What's left is shaping, cutting and checking. A human is worth more if you never write anything down and need someone to get you talking, or if you first have to figure out what you want to stand for on LinkedIn. That question is covered in How to Build a Personal Brand on LinkedIn.
How an AI employee writes in your voice
A chat window with "write me a LinkedIn post" isn't a ghostwriter yet. It needs three things: a voice profile, examples and rules for gaps.
1. The voice profile. Your tone can be written down once you break it into layers. The email version, with a table and a blind test, is in Can AI Write Emails in My Voice?. There, my email assistant's sense of my tone comes from 50 of my own real sent emails. For LinkedIn the layers shift a little:
LinkedIn voice profile
Readers: <who this is for, e.g. owners of small businesses>
Topics I write about: <three to five>
Topics I don't write about: <e.g. politics, naming clients>
First line: <e.g. always an observation, never a question>
Sentences: <e.g. short, one idea per paragraph>
Phrases I use a lot: <three to five, word for word>
Things I never write: <filler phrases, superlatives, certain emojis>
Hashtags and emojis: <none / three at most, at the end>
Stance: <where I push back, what I never promise>
2. Examples, not adjectives. Three to five of your own texts from different occasions. Emails or talk notes work if you have no posts yet. "Casual but professional" is no substitute for an example.
3. Rules for gaps. Conny's most important rule: she invents no personal facts about me, no anecdotes, no numbers, no client details. When in doubt, the spot stays blank. In your own setup, that means a missing detail shows up as a visible marker like [ADD], not as a plausible invention. The general approach to writing with AI is in Writing with AI: A Beginner's Guide.
The workflow, with approval built in
An example week with an AI employee:
- Collect material. After meetings, a two-minute voice note or a few bullet points. When you run dry, LinkedIn Post Ideas: 25 Formats and AI News helps.
- Draft. The AI turns each note into a draft that follows the voice profile, with three openings to choose from and a list of spots you need to check. If it should become a swipeable document, see How to Make a LinkedIn Carousel with AI.
- Check. Every number, name and claim against your note. A sentence that doesn't sound like you becomes a new line in the voice profile.
- Approve and schedule. You approve and schedule the post yourself in LinkedIn.
- Reply. You answer comments yourself. The AI may suggest replies.
- Learn. With Conny, reach, reactions and weak variants feed the next assignment instead of remaining isolated numbers. What didn't land belongs in the profile.
A copy-and-paste brief for a single post is in AI for LinkedIn Posts: A Step-by-Step Guide, more briefs in my Prompt Templates. Before the first posts go out, check your profile, see LinkedIn Profile Optimization.
Frequently asked questions
Is ghostwriting allowed on LinkedIn?
Neither the User Agreement nor the Professional Community Policies explicitly forbid getting help with writing. They forbid sharing your account, using bots for posts and reactions, and posting misleading information about yourself (as of November 2026). What matters is who posts, and whether the content is true.
Do I have to disclose that AI wrote my post?
LinkedIn recommends a note if you relied heavily on AI to create or modify a post and that isn't obvious from the context. Whether a legal obligation applies on top of that is covered in AI Disclosure Rules Under the EU AI Act.
Can my ghostwriter reply to comments for me?
Not from your account, which you don't share with anyone. They can suggest replies, you send them. The same goes for AI: the User Agreement prohibits commenting via bots.
Can people tell a post was written by AI?
Not reliably, detectors get it wrong too, see Can You Detect AI Text? What Detectors Do. Readers do notice generic posts, which is what the voice profile and your own material prevent.
What to do next
Fill in the voice profile above, pick three of your own texts as examples and collect raw notes for a week. Then let the AI draft the first post, check every spot and schedule it yourself. After three or four posts you'll know whether that's enough or whether you need someone who gets you talking.
Conny is available in my AI Employee Framework community as a ready-made template in Premium, with the production workflow for hooks, scripts, briefs, approval and review. Everything about it is under Community. To talk about your own setup, use the contact page.
