AI for LinkedIn posts is most useful at the drafting stage: turning your bullet points into readable text, suggesting openings, cutting length, outlining a document. You create the post itself via "Start a post" at the top of your LinkedIn homepage: write, attach a photo, video, document or poll if you like, choose who sees it, then post or schedule it. According to LinkedIn Help, a post can contain up to 3,000 characters (as of November 2026). What AI doesn't take off your hands is the substance: your opinion, your experience, your numbers and the approval. LinkedIn says so itself: you are responsible for everything you post. And its User Agreement prohibits using bots to create, comment on or like posts.
My Instagram content is drafted by an AI employee, Conny: hooks, scripts, captions. She also drafts my LinkedIn posts: short posts on AI news, each with an image that shows me as a recurring pixel-art figure. Two of her rules apply to every channel: she invents no personal facts about me, and posting is my gate. She drafts, I approve. How to approach writing with AI in general is covered in Writing with AI: A Beginner's Guide; this post is only about LinkedIn.
How to create a LinkedIn post, step by step
On desktop, LinkedIn Help describes it like this:
- Click Start a post at the top of your homepage.
- In the dropdown next to your name, choose who can see the post (Anyone, Connections only or a Group) and who can comment (Anyone, Connections only or Off). This is also where you label a brand partnership.
- Type your text in the "What do you want to talk about?" field. The icons below let you add photos, videos, an event, a document or a poll.
- Click Post, or pick a time via the clock icon.
In the app, tap Post in the navigation bar. Drafts can be saved and finished later, even on another device. According to LinkedIn Help, you can schedule a post between 10 minutes and 3 months ahead; events and jobs can't be scheduled.
The same help page also says: "If we detect excessive post creation that indicates inauthentic activity, or the use of an automation tool, we may limit the visibility of those posts." More posts don't automatically mean more readers.
Formats: text, image, document, video, poll
All figures as of November 2026. On documents, LinkedIn Help adds: anyone who can see the post can download the document as a PDF, the document can't be changed after uploading (only the description can), and videos or animations inside it show as static images. LinkedIn recommends converting to PDF, and PDFs with pages of different sizes must be fit to one page size.
For video, LinkedIn advises keeping the edges free of key elements, text and logos so the interface doesn't cover them. How to create subtitles is covered in Auto-Generate Subtitles: A Guide.
The posting help page names no fixed point at which the feed collapses your text, so make your first one or two sentences work on their own.
Using AI for LinkedIn posts in five steps
A workflow, as an example:
- Write the raw note yourself. Five to eight lines: what happened, what you learned, what the reader should take away. Conny works the same way for my Instagram captions: she doesn't get the finished reel, she gets the observation note, the same five to eight raw lines the hook and the script come from.
- Brief the AI with examples. Two or three of your own posts as a model for tone do more than the instruction "casual but professional".
- Ask for several openings. Three versions to pick from are faster than three rounds of revisions.
- Check. Every number, every name, every claim against your note.
- Post it yourself.
A brief to copy:
Goal: A LinkedIn post in English, addressing the reader as "you",
about <length> characters.
Audience: <for example owners of small trade businesses>.
My tone (two posts of mine): <paste here>
Material (my raw note): <five to eight lines: what happened,
what I learned, what the reader should do>
Task:
1. Suggest three different first sentences, each clear on its own.
2. Write the post using the opening that fits the material best.
3. Hashtags: <none / at most three at the end>.
Limits:
- No experiences, numbers, clients or quotes that aren't in my raw
note. If something is missing, mark the spot with [ADD].
- No filler like "In today's fast-paced world".
- You don't publish anything; the result is a draft.
Format: the three openings, then the post, then a list
"Spots I need to check".
You set the length; LinkedIn allows up to 3,000 characters. The [ADD] marker is the same idea as Conny's rule: if a detail is missing, the draft shows a visible gap instead of a plausible invention. A more general version for blog posts, newsletters or LinkedIn is template 8 in my Prompt Templates.
For a document post, change the task, for example: "Turn my guide into an outline for an eight-page PDF, each page with a short heading and one or two sentences." Canva or PowerPoint handle the design; for your own images see Creating AI Images: Free and Free to Use. If the same character should appear in all your posts, the post on creating an AI character shows how to keep it consistent.
What AI shouldn't do for you
Your opinion and your experience. According to LinkedIn, AI tools can support "refining language, proofreading or being more concise", but "content should still reflect the member's own voice, perspective, and experience." LinkedIn calls generic AI content without a point of view "AI slop", and says content that is "generic, repetitive, or lacks a clear point of view is less likely to be widely distributed."
The approval. In LinkedIn's words: "We urge you to review, edit, and approve of any content that you've used AI to assist with. You are ultimately responsible for everything you post on LinkedIn". My own LinkedIn posts work the same way: Conny drafts, I approve.
Posting, commenting, liking. LinkedIn's User Agreement prohibits using "bots or other unauthorized automated methods" to, among other things, "create, comment on, like, share, or re-share posts". AI may suggest a reply to a comment. You send it yourself.
Other people's rights. AI-assisted content must not infringe others' rights, LinkedIn says, "including intellectual property and privacy rights." For AI images, see AI and Copyright: Who Owns AI Images?.
Disclosing AI use. LinkedIn recommends letting others know if you've relied heavily on AI to create or modify a post, unless that's obvious from the context. Whether the law requires a label on top of that in the EU is explained in AI Disclosure Rules Under the EU AI Act. That's my operator's perspective, not legal advice.
One piece of content, two channels: LinkedIn and Instagram
An example, made up: a painting business has finished a facade using a new method. From one raw note come a short LinkedIn post with the lesson for other businesses, a five-page PDF on the process and, for Instagram, a caption for the before-and-after photo. In this example, LinkedIn reaches partners and job applicants, Instagram private customers. So give the AI a separate brief per channel instead of the finished LinkedIn post to rewrite.
How a caption is built is covered in How to Write Instagram Captions. The Instagram counterpart to the LinkedIn document is the carousel, see Instagram Carousel: Size, Slides, Structure. If you also publish a video as a reel, you can try it out first as a trial reel, see Instagram Trial Reels: Setup, Results, Limits. LinkedIn as a source of business contacts is a separate topic, covered in Finding B2B Leads: LinkedIn and Alternatives.
Frequently asked questions
How long can a LinkedIn post be?
Up to 3,000 characters, according to LinkedIn Help (as of November 2026). Longer texts go into articles, which you write on desktop only.
Can I schedule LinkedIn posts?
Yes, via the clock icon in the post window, between 10 minutes and 3 months ahead. Events and jobs can't be scheduled.
Is LinkedIn's AI post generator still available?
Not at the moment. LinkedIn's help page on its AI-powered writing tool is titled "No longer available" and says new tooling will come "in the near future" (as of November 2026). For drafting, use ChatGPT, Claude or Gemini.
Does LinkedIn show AI posts to fewer people?
LinkedIn says its focus "is not on how content is created, but whether it adds value." Members can flag posts via the menu as "Seems like AI slop". If a post gets enough of that feedback, you'll see a tip in your post analytics. According to LinkedIn, that does not mean your post was taken down.
How do I see whether a post works?
Click the Me icon, then Posts & Activity under Manage, and on a post click the number of impressions or View analytics. You see, among other things, reach inside and outside your network, engagement and viewer demographics.
What to do next
Take your next LinkedIn idea and write the raw note first: five to eight lines about what actually happened. Give it to an AI chat with two of your own posts and the brief above, pick an opening, fill every [ADD] yourself and post only when every sentence is true. After a few posts, your analytics show which format lands.
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.
