LinkedIn post ideas come from five sources, four of them already in your business: the questions clients keep asking, your own numbers, your mistakes and your industry's news. The fifth is AI news: new tools, prices and rules that affect your readers. It never runs dry, so you need a filter: only what fits your positioning, and where you have something of your own to add, becomes a post. Below: 25 formats with an opening line to fill in, a workflow in which an AI employee screens the news daily and suggests posts with a hook, and a prompt to copy. Choosing and approving stays with you.
My LinkedIn posts are drafted by Conny, my AI employee for content: short posts on AI news, each with an image that shows me as a recurring pixel-art figure. I approve every post myself. How a single post comes together is in AI for LinkedIn Posts: A Step-by-Step Guide; everything on the topic is on LinkedIn with AI. This article covers the step before: what to write about.
Where LinkedIn post ideas come from: five sources
Source
How you spot the moment
What you note right away
Client questions
You give last week's answer again
The question verbatim, no names, your answer in one sentence
Your own numbers
A report surprises you
The number, the period, what you expected
Mistakes
You clean up something that went wrong
What happened, what you do differently now
Industry news
A client asks you about a headline
The link, who it matters to, your take
AI news
A tool your readers use changes
What changes, from when, what your readers need to do
The first four work for Instagram too; how to collect them with a raw note is in Finding Reel Ideas for Your Business. On LinkedIn, readers want to know what it means for their work.
Why AI news deserves its own slot
There is always a new occasion: a new model, a renamed plan, a changed limit. Almost every reader then asks: what does this mean for me? Others have shared the headline; what's missing is the context for your people.
In its guidance on AI-assisted content, LinkedIn writes that its "focus is not on how content is created, but whether it adds value", and that content "should still reflect the member's own voice, perspective, and experience" (as of November 2026). A retold press release does neither; "for your firm, this changes exactly one thing" does both. The risk is volume: comment on everything and you become a news ticker.
How an AI employee screens the news daily and suggests posts
Screening repeats, follows fixed rules and ends with a suggestion, not a published post. That suits an AI employee. An example workflow:
Pick your sources. Few and reliable: news and changelog pages of the vendors your clients use, two trade publications, one newsletter you read anyway. Primary sources first.
Screen once a day. Only what's new since the last run: headline, link, date.
Filter: does it fit my positioning? Three written questions: Does it affect my audience? Can I add something the headline doesn't say? Is it confirmed at the source? One no, and it's out.
Suggest. Per item: two or three opening lines, the core in one sentence, who it matters to, a format from the list below and the link. Three to five a day are plenty.
A human picks and approves. You take one suggestion or none and add your opinion and experience; only then is a draft written. You post it yourself.
That step isn't politeness. LinkedIn's User Agreement says you may not "use bots or other unauthorized automated methods" to, among other things, "create, comment on, like, share, or re-share posts" (effective November 3, 2025). The AI screens and drafts; the click on "Post" is yours.
In my business, Conny writes the drafts. I give her a topic, a link or a longer recording; she turns it into openings and copy in my voice. Two of her rules fit news posts: she invents no personal facts about me (when in doubt, the spot stays blank), and posting is my gate: she drafts, I approve.
25 LinkedIn post ideas as templates
Each idea is a format with an opening line. Fill the angle brackets with your material, not with the AI's plausible guesses.
From client questions
Question of the week: "I heard this question times this week: <question, word for word>."
The short answer: "The short answer to is . The long one is below."
The wrong question: "People ask me . The better question is ."
Before the first call: "Three things to sort out before you call a ."
The decision: "
From your own numbers
The surprising number: "I expected . It was ."
Before and after: " ago: . Today: . Here's what changed in between."
The time math: " now takes
The monthly review: "My in three numbers, and what I'm doing about them."
From mistakes
My mistake: "I made . It cost me ."
The fix: "I find this with almost every new client: ."
The rule I dropped: "I did for years. Not since ."
The checklist after: "Since , I check these three things."
From industry news
What it means for you: "From , applies. For , that means ."
The counterpoint: "Everyone is talking about . I think it's , for one reason."
Three consequences: ". Three consequences the coverage misses."
The deadline: "By , you need to . Here's how to approach it."
From AI news
The change in five lines: " has . What it means for , in five lines."
New, but for whom? " is new. It pays off for , not for ."
The hands-on test: "I tried on . Result: ."
The price in plain terms: " now costs . Here's the math you need to do."
The limit: "What can't do after the change: ."
What I'm changing now: "Because of , I'm switching . And this stays the same."
From your stance
What I don't hand off: "This is a task I deliberately keep away from AI: ."
A question for your network: "I'm facing . How did you handle it?"
Numbers 20 and 21 need evidence: a test only if you ran it, a price only with the date from the vendor page. Many formats also work as a swipeable document. How to build one is in How to Make a LinkedIn Carousel with AI. For formats 1 to 5, template 7 in my Prompt Templates turns real client questions into answers.
Prompt to copy: screen the news and suggest posts
Once a day in an AI chat, with the new links or texts:
Role: You screen news for my LinkedIn posts.
My positioning: <one sentence: who I help with what>
My readers: <for example, owners of small accounting firms>
Topics I don't write about: <for example, politics>
Material: <links or texts of the news since yesterday>
Task:
1. Check each item:
a) Does it directly affect my readers?
b) Can I say something that isn't in the item itself?
c) Is it confirmed at the primary source?
Any no: drop it, reason in a few words.
2. For each remaining item, five at most:
- three opening lines, each clear on its own
- the core in one sentence: what changes for my readers?
- a format (change in five lines, counterpoint,
deadline, limit, price in plain terms)
- the link to the primary source
3. Sort by usefulness to my readers.
Limits:
- Don't invent any opinion, experience or number of mine.
Where my take is missing: [MY TAKE].
- No prices or deadlines without a link to the source.
- Suggestions only. Publish nothing.
Review and approval. LinkedIn writes: "We urge you to review, edit, and approve of any content that you've used AI to assist with," and "You are ultimately responsible for everything you post on LinkedIn." With news: check every detail against the source, not the AI's summary.
Other people's content. Link the item and summarize it in your own words rather than copying text or images. Whether to disclose AI assistance is covered in AI Disclosure Rules Under the EU AI Act. Neither is legal advice. Where the line is when an AI writes whole posts for you is in AI as Your LinkedIn Ghostwriter.
Frequently asked questions
What should I post on LinkedIn when I'm out of ideas?
Answer the last question a client asked you in five lines (format 1 or 2). Others have it too and never ask.
Can AI come up with LinkedIn post ideas?
Yes, from material: news, notes, a recording. Without it, you get the generic ideas from every list.
How do I find AI news that fits my industry?
Read the news or changelog pages of the tools your clients use, plus two trade publications. The filter matters more than volume.
What to do next
Today, pick three sources and write the filter questions above them. Tomorrow, paste the new items with the prompt into an AI chat, take the best suggestion and add the sentence the headline lacks. Collect client questions too, so not every post is news. How an AI employee handles this is on Conny's page. How I run my AI employees is in my newsletter "From the Engine Room": pick a resource on my free templates page and sign up. Before new readers click through to your profile, see LinkedIn Profile Optimization: Tools and AI. If you mainly want client contacts from LinkedIn, read Finding B2B Leads: LinkedIn and Alternatives.
Kevin Welter
Developer, IT architect, author of technical books (Kubernetes, cloud infrastructures) and speaker. Runs his business with an AI workforce of fifteen AI employees and shows solo business owners in his community how to hire their first AI employee.
