Before you post a comment on LinkedIn, ask one question: would it make sense under a different post? If it would, it is a template, the kind of comment LinkedIn said in July 2026 it is building systems to keep from spreading widely.17 I call it the swap test.
This is one of four deep dives behind my essay on LinkedIn's AI slop crackdown.
In short
- Run the swap test. A comment that would fit under someone else's post adds nothing, and that is the quality failure LinkedIn names.
- The risky draft is the polished one that restates the post. Add a case, a cost, a number, or the conditions where the advice breaks.
- A weak comment stays up and sinks. LinkedIn ranks comments for each reader and says it will not broadly distribute ones that add nothing.
- Commenting fast can get your comments limited even when no tool is involved.
A weak comment stays up and sinks
Rishi Jobanputra of LinkedIn described the comments it is building systems to "not broadly distribute": "comments lacking meaningful substantive input," "comments that repeat what's already in the post without adding new ideas," and "responses that provide little insight or value to the conversation."17
LinkedIn does not say it deletes them. Comments are ranked for each reader, using "your professional interests, connections, and engagement activity,"17 and the default view is "Most relevant."11 The Help page says, without naming which, that "Some comments will only be visible under Most recent view."6 I suspect few readers switch views, so those comments are close to invisible, and your own view of a thread proves nothing about anyone else's.
Racing to comment first matters less than people think. In 2017 LinkedIn's engineers wrote that ranking comments by likes left "early comments on a thread having an unfair advantage," and built a personalized ranker instead.63 It still weighed freshness, so early helps a little, but a fast "Great post!" can no longer bank likes ahead of everyone. A specific comment from a stranger can be among the first things your buyer sees.
No LinkedIn statement I read names a banned phrase or a minimum length, and AI help is allowed as long as the comment carries your voice and perspective.9 Anyone selling a list of forbidden comment phrases is guessing.
The polished draft is the one that fools you
Nobody needs to be told that "Great insights! Thanks for sharing!" is weak. The comment that fools people, including the person posting it, is the polished one. It names the author, sounds engaged and ends with a question. LinkedIn says its systems are trained to separate writing that "adds perspective, context, or expertise" from writing that "feels generic or repetitive, even if it appears polished on the surface."9
Before
Love this, Priya. Moving away from per-seat pricing makes so much sense as buyers push back on paying for every login. Curious how you landed on the right number for the flat fee?
After
A 35% jump in deal size is a strong two quarters. The number I'd watch next is expansion. When we went flat, we lost seat growth as our upsell trigger, and net retention dipped a year later. Did you add a usage tier?
Illustration; the commenter is made up. The before praises, restates and asks what anyone would ask, so it fits a hundred posts. The after only makes sense under this one, and it hands Priya a cost she skipped.
Add one thing the post does not say
In 2015 LinkedIn's sales blog told reps to "provide additional insight from your own experiences in the industry" and added, "This is not the time to sell your solution, however."64 What is new is that LinkedIn is building systems to hold back comments without it.17 Pick one move:
- A counter-example. A case where the same move went the other way.
- A cost the author skipped. The risk that shows up a quarter later.
- Your own number. A figure from your work.
- The conditions. When the advice holds and when it breaks.
- Math on the author's numbers. Their own figures often hide a consequence they did not spell out.
- The buyer's side. If you sell, you know which objection a prospect would raise. Say it.
The last two need no data of your own, and no data is when most people reach for ChatGPT filler.
"I agree" gives the author nothing to answer. "I agree, and it only held for us on deals under 50 seats" does.
Before
So true. Flat pricing beats per-seat every time. Simplicity sells.
After
Haven't tried this, and approvals are what would stop us. At the mid-size companies we sell to, a contract 35% bigger often crosses a sign-off limit and adds a finance review. Did your sales cycle get longer on the bigger deals?
Illustration; the commenter is made up. No outside data. The after uses the post's own number and one fact about the commenter's buyers.
If you cannot write something real, skip the post.
Ask ChatGPT for questions, never for your experience
Asked to "write a comment," a chatbot restates the post. Asked what you "might add," it invents experience you do not have. Ask it for the open questions and write the comment yourself:
Answer one from your own work, in your own words, then run this list.
Save this
Before you post a LinkedIn comment
- Would it make sense under a different post? Rewrite it.
- Delete the first sentence if it restates the post.
- Cut the praise. A word that only compliments goes.
- Add one thing the post lacks: a case, a cost, a number, or a condition.
- Find the one claim you would defend on a call. No claim, no comment.
- Read it aloud. If you would not say it to the author's face, rewrite it.
- Paste it and click Comment yourself, a few a day.
handful.cc/blog/linkedin-comments
Paste from a chatbot in its own tab. An extension that drafts inside LinkedIn's page is banned even when you click Comment yourself, for reasons in the piece on AI tools.
Speed alone can get your comments limited
LinkedIn publishes no number. Its Help page says it "may limit how many comments a member or a LinkedIn Page can make in a certain time period," and that "if we detect excessive comment creation or use of an automation tool, we may limit the visibility of those comments."6 That "or" means volume alone can be enough.
Five comments a day that each say something are worth more than fifty that say nothing, and fifty in an afternoon is the kind of pattern that looks like a tool.
Finding the post is the other half
LinkedIn's executive editor put the upside well:
“A single insight shared at the right time can spark a conversation, open a door, or create a new opportunity.”
"At the right time" is the hard part. I sell for two companies, and when I searched LinkedIn by keyword I got competitors, job ads and students. The one good post was usually a day old by the time I found it.
So I built Handful. It finds up to five LinkedIn and Reddit posts a day from people who need what you sell, never competitors, and drafts a comment in your voice that responds to what the person said, never "Great insights!" You post it yourself. Handful never logs in or touches your account. Five is a ceiling, so on a thin day you get fewer.
Next in the series: LinkedIn Bans the AI Extension Even When You Click Comment Yourself.
Sources
- 06LinkedInComment on posts and reply to a comment, LinkedIn Help Center, ~2026-07. linkedin.com
- 09LinkedInKeeping conversations real on LinkedIn, Laura Lorenzetti, LinkedIn Pulse, 2026-05-20. linkedin.com
- 11LinkedInFriday update: automated comments, Gyanda Sachdeva, VP Product, LinkedIn post, 2026-02-13. linkedin.com
- 17LinkedInMaking LinkedIn conversations more relevant and valuable, Rishi Jobanputra, LinkedIn Pulse, 2026-07-31. linkedin.com
- 63LinkedInServing Top Comments in Professional Social Networks, LinkedIn Engineering, 2017-09-20. linkedin.com
- 64LinkedInHow to Use Comments to Start Conversations with Prospects, LinkedIn Sales Blog, 2015-01-24. linkedin.com