Social Media, AI Marketing
•
7 min.

by Olivia Ulbing-Sommeregger

In a nutshell
LinkedIn classifies suspected AI-generated comments as less relevant; in tests, their views dropped by more than 20 percent.
The comments are not deleted, but they are sorted downward in the default view and trigger notifications less frequently.
AI itself remains permitted; text lacking personal experience or perspective is what gets downgraded.
According to Meltwater, LinkedIn is the second most cited source in AI responses, with around 75 percent coming from personal profiles.
For you, this means shutting down comment bots, building two to three authentic voices, and checking AI visibility monthly.
✓ Do’s
Comment with substance
Your own figure, a practical example, or a counter-question. Three sentences backed by experience beat "Great post!" ten times over.Target and build a few voices
Two to three people from management and specialized departments who post on clear topics achieve more than a mandatory post for everyone.Use AI as a sparring partner
Get suggestions for counterarguments or a structure. The wording and opinion remain yours.
✕ Don’ts
No Auto-Comment Tools
Extensions that post comments automatically violate LinkedIn's rules and now end up noticeably further down.No Engagement Pods
Obligatory comments among friends are a pattern, not a conversation for LinkedIn. Anyone counting this as a KPI is measuring the wrong metric.Don't Just Let the Company Page Do the Talking
Accounting for around 25 percent of LinkedIn citations in AI responses, the company page represents only a quarter of your visibility.
How do we measure if this works?
From now on, the number of comments alone is a shaky KPI. If LinkedIn devalues weak comments, the number might drop while quality increases. At oha!, we therefore look at three things: genuine replies to comments, profile views and inquiries referencing posts, and mentions in AI responses.
For the last point, a simple list is enough to start with. Note down ten questions that your clients would type into ChatGPT, Gemini, Claude, and the like, and check monthly who is mentioned and which source is linked. It's not rocket science, but it's a starting point you can report upwards. We show how we systematically approach AI visibility in AI marketing.
What changes in day-to-day social media?
Commenting becomes manual work again, and that's a good thing. Schedule a fixed 15 minutes a day for five genuine comments on posts from clients, partners, and industry voices—ideally from the people who need to become visible. In the editorial plan, each post is assigned one of three or four core topics, just as LinkedIn itself recommends. Ghostwriting remains an option, as long as the opinion and experience come from the person whose name is on it.
That's how we at oha! are currently setting this up in social media management for our clients. We gather thoughts in a brief interview, take notes, and the person decides on every single word. This takes longer than a prompt, but delivers content that LinkedIn and AI searches recognize as human. And yes, the same applies to Instagram, where follower counts lie anyway.
Today
Shut down bots
Check which tools, extensions, or agencies are commenting automatically on your behalf, and turn them off.
This week
Start a commenting routine
15 minutes a day, five comments with your own figure, experience, or counter-question. Track replies and profile views.
This month
Define topics and voices
Define three or four core topics and select two to three people who can speak credibly about them.
This quarter
Measure AI visibility
Test ten customer questions monthly in ChatGPT, Gemini, Claude, and others, and show the progress in reporting.
Is AI now banned on LinkedIn?
No. LinkedIn itself says that AI can assist with writing, as long as the post and comment reflect your voice and your perspective. It's the shortcut that gets penalised, not the tool. The platform rewards what has always worked: people with opinions responding to other people.
In the interest of full transparency, we make money from social media management and LinkedIn support. So you are welcome to be sceptical when we say the effort is worth it. But the figures above come from LinkedIn and Meltwater, and they have their own interests too: LinkedIn wants more posts, Meltwater sells tools for AI visibility. Our advice remains simple. Turn off the comment bots and let your best minds speak for themselves.
Are you mentioned in ChatGPT?
Let’s talk
Who speaks for your company on LinkedIn?
More on social media management
→
FAQ
Frequently Asked Questions about AI Comments on LinkedIn
Is it worthwhile for our management to be personally active on LinkedIn?
If you want to appear in AI searches, yes. According to Meltwater, around 75 percent of LinkedIn citations in AI responses come from personal profiles and only 25 percent from company pages. One to two posts per week sharing real experience are enough to start; a large following is not necessary for this.
What do we risk if our team uses AI tools for comments?
Visibility and credibility. Automated comments violate LinkedIn's rules and are now penalized; in tests, their views dropped by more than 20 percent. On top of that comes reputational damage if customers spot the bot.
How do we measure whether LinkedIn is making us more visible in ChatGPT and other platforms?
Define ten typical customer questions and check them monthly in ChatGPT, Gemini, Claude, and similar tools. Note whether you are mentioned and which sources are linked. Complement this with profile views, replies to comments, and inquiries referencing posts.
Do we need to shift budget from the company page to individuals?
More time than money. The company page remains important for ads and employer branding. In addition, it is worthwhile to allocate resources for two to three people, such as for interviews, ghostwriting, and a dedicated commenting routine.
Can I still use AI for LinkedIn comments?
As an aid, yes; as the author, no. LinkedIn permits AI for drafting as long as the opinion and experience come from you. Have it suggest counterarguments, but write the comment yourself and avoid platitudes that merely repeat the post.
What can I specifically test this week?
Set aside 15 minutes a day for five comments with your own data point, your own example, or a follow-up question. After two weeks, compare replies, profile views, and new connections with the weeks before.
Sources
Hari Srinivasan (LinkedIn CPO) on LinkedIn: Update on AI slop comments (November 2026)
Social Media Today: LinkedIn cracks down on AI-generated comments (15/11/2026)
Social Media Today: LinkedIn says 1M people have reported AI slop (20/08/2026)
Campaign: LinkedIn cracks down on automated content with new ‘seems like AI slop’ detection button
Social Media Today: LinkedIn wants to limit the reach of AI-generated content
Social Media Today: LinkedIn outlines more measures to combat engagement pods

Written by
Founder & CEO, oha! digital
Olivia Ulbing-Sommeregger is founder and CEO of oha! digital, an agency for social media, performance marketing and AI based in Salzburg. Since 2018 she has been supporting companies across Austria, Germany and Switzerland – from SMEs to brands like Bankhaus Spängler and Miele Professional – with strategy, content and advertising on Meta, TikTok, LinkedIn and Google. She is co-author of the Rheinwerk book “Follow me! Erfolgreiches Social Media Marketing” (7th edition, 2025) and completed the WIFI diploma programme in CSR, ESG and sustainability management. Her standard is straight talk instead of agency blah.
← All Klartext articles
Keep reading
More Klartext
All articles →

Social Media, Performance Marketing
•
Social media ban in Austria: Brussels hits the brakes – your budget doesn't
→

AI, social media
•
Meta Muse Austria: The AI agent no one has (yet) – but everyone should know about
→

Social media, content
•
Instagram Features: Mosseri Was Against Long Video – Now 90 Minutes Are Coming
→
Ready when you are!
Let’s make your
digital story work.
Tell us briefly what you’re planning.
The first call is on us.
