Course

Clients from LinkedIn comments

How to get clients by commenting on other people’s LinkedIn posts: find the right posts without scraping, write a comment that earns a reply, make the free ask, and turn “add me” into a customer. Fact-checked against LinkedIn’s own rules.

Lesson 1

What this is, and who it is for

Most people who sell a service to other businesses get clients one of three ways: they post content and wait, they send cold messages, or they pay for ads. All three work eventually. All three are slow or expensive at the start, when you have no audience and no budget.

There is a fourth channel that costs nothing and needs no audience of your own: commenting on other people’s posts. When your buyer is reading a post about the problem you solve, and the best comment under it is yours, you have been introduced to that buyer by the person they already follow. That is what this course teaches, end to end, from choosing the offer to counting the clients.

It is for anyone who sells to businesses on LinkedIn: agencies, freelancers, consultants, software founders, and the Shopify operators and virtual assistants this site mostly writes for. It assumes you have a LinkedIn profile that says what you do. It does not assume followers, a newsletter, or any paid tool.

Where it came from

The method is from an eight-minute video by Amit Yadav, published on 8 September 2026, which we read in full and checked line by line before writing this. The core of it holds up and is credited. Two parts of it break LinkedIn’s rules, and this course replaces them rather than repeating them. Where that happens, you will see a block marked Checked saying exactly what the source said, what is true, and what we changed. There are eight of those. Read them; they are the difference between a method that works for a year and one that gets your account restricted in a week.

Lesson 2

Decide the offer before you write a word

Every comment in this method ends with an ask. The ask has to be decided before you look at a single post, because it shapes which posts you look for and what you say under them.

The ask is free. That is the rule, and the reason is trust.

Someone reading a stranger’s comment has given you about four seconds and no goodwill. A comment that ends in “book a call” or “here is my pricing” asks for money from a person who has not yet decided you know anything. A comment that ends in “I keep a sheet of the twelve subject lines that beat 60 percent open rates, want it?” asks for nothing except permission to be useful. The second one gets replies. The first one gets scrolled past.

The two halves

  • The front end, free. A small, specific thing you can hand over within a day of someone saying yes. A comparison sheet, a checklist, a teardown, a template, a private group where people compare notes on one narrow topic. It has to be genuinely useful on its own, or the person who asked for it learns you waste their time.
  • The back end, paid. What you actually sell. The front end should sit naturally in front of it: the person who wanted the subject-line sheet is a person who might pay for someone to run their email. You never mention the back end in a comment. It comes up later, in conversation, when they ask.

The one exception

If a post is literally asking for a vendor (“anyone know a good Klaviyo freelancer?”), answer the question. Pitching the paid thing is what was asked for. Everywhere else, the free ask.

Checked The free-first rule is the strongest part of the source, and it is right.

The video says “CTA should be ideally a free offer” and “if you are adding anything paid it is going to be very difficult for you to get them to trust you,” and makes the same exception for posts that are already asking for a tool or an agency. We kept all of it. This matches how any cold channel works: the first exchange has to cost the other person nothing.

Write it down

One page, kept open while you comment. Who you sell to, in one sentence. What they are struggling with, in five bullets. The free front end, in one sentence, with the link or the group ready. The paid back end, in one sentence, so you remember what all this is for. You will paste this page into an AI later, in lesson five, so write it as if explaining to a new hire.

Lesson 3

Find the posts, without scraping

The posts you want are the ones your buyer is reading and reacting to, about the problem you solve, published recently enough that a comment still gets seen. LinkedIn gives you three ways to find them that do not break its rules, and they take about twenty minutes a day.

1. Search posts, filtered to this week

Type a phrase your buyer would use into the LinkedIn search bar, press enter, then choose Posts and set Date posted to past week. Not your jargon, theirs: a Shopify owner does not search “retention strategy,” they write “abandoned cart” and “repeat customers.” Try five or six phrases. Save the ones that return real conversations rather than link dumps. The results are sorted by engagement by default, which is what you want: a post with forty comments has forty people reading the comments.

2. Follow the thirty accounts your buyers follow

Every niche has twenty or thirty people whose posts your buyers read. Find them by looking at who your existing clients follow, and who is commenting thoughtfully on the posts from step one. Follow them, and use the bell on their profile to be notified when they post. A comment in the first hour of a post from one of these accounts is worth ten comments on dead posts, because the author’s audience is arriving at the same time you are.

3. Read the comment sections

The comments under a busy post are a list of people with the problem, in their own words, today. Some of them are better targets than the post itself. Reply to those too; a reply to a comment notifies its author directly.

The daily routine

  1. Open notifications. Comment on any post from your thirty that is under two hours old.
  2. Run your saved searches, past week. Pick five posts with real discussion under them.
  3. Read each post properly. Read the comments. Then write.
  4. Log every comment in the sheet from lesson eight before you move on.

Checked The source finds posts by scraping five hundred of them. That breaks LinkedIn’s rules, and we do not teach it.

In the video the posts come from a prompt to Claude Code: “scrape me 500 LinkedIn posts where people are talking about GTM and cold outbound.” LinkedIn’s User Agreement, section 8.2, lists what members agree not to do. Item 2, verbatim: “Develop, support or use software, devices, scripts, robots or any other means or processes (such as crawlers, browser plugins and add-ons or any other technology) to scrape or copy the Services, including profiles and other data from the Services.” Scraping from a logged-in account is the thing that clause names, and LinkedIn restricts accounts for it. The search routine above finds the same posts, at the pace you can actually comment on them, with nothing to explain later.

Lesson 4

Write a comment that earns a reply

A comment has one job: to make the post’s author and the people reading it think “this person actually knows this.” Everything about the structure follows from that.

The shape

  1. One or two paragraphs of specific value. Not agreement. Not a compliment. Something the post did not say that the reader is glad to know: a number from your own work, a step the author skipped, a case where the advice fails, a sharper way to do the thing described. It must be obvious you read the post.
  2. Then the free ask, tied to that post. One sentence. It offers the front end from lesson two, phrased in the terms of the post you are under, and it ends in a question so the reply is easy: “want it?”, “happy to add you”, “shall I send it over?”

The specificity test

Read your comment back and ask: could this be pasted under a different post about the same topic? If yes, it is not specific enough and it will read as what it is. The ask in particular should be different under every post, because it names the thing that post was about.

Under a post comparing LinkedIn message formats

The one variable that moved our reply rate more than format was whether the first line mentioned something from the person’s last post. Same template otherwise, 4 percent to 11 percent over about 300 sends. Format mattered less than we expected.

I keep a sheet where a few of us compare format tests like this one against actual reply rates. Want me to add you?

Under a post about SDR qualification questions

Question three on your list is the one that gets skipped when the call is going well, and it is the one that predicts whether the deal closes. We started scoring calls on whether it was asked at all, and the ones where it was closed at roughly twice the rate.

There is a small group where SDRs compare qualification questions like these against booked-to-held numbers. Happy to add you if useful.

What not to write

Great post! Totally agree, consistency is key. I help agencies scale with done-for-you outbound, DM me if you want to chat.

The bad one fails every test. It could sit under any post. It offers nothing. It asks for a sales conversation from a stranger. And it is the comment LinkedIn’s readers have learned to skip, so it damages the reader’s opinion of the four good comments you wrote earlier.

Small rules

  • No link in the comment. Make the ask, and send the link when they say yes. A comment that is only a link is a comment nobody reads.
  • Under 120 words. Two paragraphs and a question. The post is the main event; you are a footnote worth reading.
  • Never the same ask twice on the same day. If your front end is one thing, describe it differently every time, in the post’s terms.
  • If you cannot find something specific to add, do not comment. A day with four real comments beats a day with fifteen empty ones.

Checked The comment structure is the source’s, and it is the reason the method works.

The video describes exactly this: “the first two paragraphs are just about the value and the last one is the call to action, which has to be free,” and it shows the ask changing under each post (“we are not pitching a generic thing, we are pitching a very specific thing to that particular person”). The two good examples above are rewritten from the ones on screen in the video. The numbers in them are illustrative, and yours should be real.

Lesson 5

Use AI to draft, and a person to post

Writing five specific comments a day, every day, is where most people quit. An AI model is good at the part that makes it slow: reading a long post carefully and producing a first draft that engages with what it actually says. It is bad at the part that makes it work: knowing what is true about your business, and knowing when a draft sounds like a machine.

The prompt

Paste your one-page offer document from lesson two. Then the rules from lesson four, in full: two short paragraphs of specific value, a free ask tied to this post, under 120 words, no link, no compliments, no “great post.” Then the post itself, copied from the page. Ask for three drafts, not one, so you are choosing rather than accepting.

What you do with the draft

  1. Check every fact. The model will invent a plausible number if you let it. Replace it with yours, or cut it.
  2. Cut the first sentence if it restates the post. It usually does.
  3. Rewrite the ask in your own words. It is the sentence people reply to; it has to sound like you.
  4. Post it yourself, from your own account, by hand.

Checked The source has the AI write the comments at scale, and stops short of saying who posts them. Automated posting breaks the rules; identical text is what gets caught.

In the video Claude Code reads each scraped post and writes the comment, five hundred at a time, into a sheet. Section 8.2 item 13 of LinkedIn’s User Agreement: do not “use bots or other unauthorized automated methods to access the Services, add or download contacts, send or redirect messages, create, comment on, like, share, or re-share posts, or otherwise drive inauthentic engagement.” The Professional Community Policies separately forbid “gratuitously repetitive” content. Generating a draft is neither of those. Posting it by machine is the first, and posting the same opening line under thirty posts is the second. Keep the AI on the drafting side of that line.

Lesson 6

How many, how fast, and what gets accounts restricted

LinkedIn publishes no daily comment limit. What it does publish is the behaviour it punishes: automation, repetition, and anything that looks like an unreasonable load. Everything below is about staying on the right side of that while doing enough to matter.

The numbers

  • Start at five good comments a day. Not five minutes; five comments that pass the specificity test. That is about twenty minutes once the searches are saved.
  • Grow to ten or fifteen over a month if replies are coming. Past that, quality drops before the numbers improve, and the people who see three comments from you under three posts in one afternoon start to notice the pattern rather than the content.
  • Spread them across the day. Fifteen comments in eight minutes is a burst, and bursts are what filters look for.

What a restriction looks like

Comments that post but no one can see, then a notice that your account is restricted. If it happens: stop, disconnect any browser extension or tool that touches LinkedIn, leave it for two days, and come back doing everything by hand at half the previous volume.

Checked The source gives no volume guidance and no numbers. Reports of a comment cap are practitioner accounts, not a LinkedIn statement.

The video generates five hundred drafts and says reply rates are “really, really high” without a figure. In spring 2025 a number of LinkedIn users reported comments being throttled at around thirty a day; LinkedIn has never confirmed a limit. What it has written down is the User Agreement clause against placing “an unreasonable load on the Services (e.g., spam...).” The figures in this lesson are ours, chosen to stay well under any reported threshold; treat them as a starting point, not a rule.

Lesson 7

From “add me” to a client

A reply that says “would love to join” or “feel free to add me” is the whole point of the previous four lessons, and it is where most people fumble. The person has said yes to something free, in public, and they expect it to arrive.

The first hour

  1. Reply under the comment, briefly: “Sending it now.” That is visible to everyone else reading, and it shows the ask was real.
  2. Send a connection request with a one-line note that names the post, so they know who you are when it lands.
  3. The moment it is accepted, send the thing. The link, the file, the group invite. No preamble, no questions about their business yet.

The next two weeks

Now they are on your list, in your group, or holding your sheet. The job is to be useful again, twice, without asking for anything. If it is a group, that means your own posts in it answer questions people actually asked. If it is a document, that means one follow-up message a week later: “did the sheet help? one thing I would add is...” People buy from the person who was useful three times before asking.

The paid conversation

It starts when they ask, and they ask in one of two ways: a direct question (“do you do this for clients?”) or a problem stated with no question mark (“we tried this and it did not work”). Both are the opening. Answer the question, then offer a short call to look at their specific situation. The call is where the back end from lesson two is sold; nothing before it is.

If you are doing this for someone else

If the group or the sheet belongs to a client and you are being paid to promote it, say so in the comment, in three words: “I work with a group where...” rather than “I’m part of a group where...” Readers in the United States are covered by the FTC’s endorsement rules, which require a paid connection to be disclosed, and readers everywhere can tell the difference between a recommendation and an advert once they find out. The disclosure costs almost no replies. Getting found out costs all of them.

Checked The source is promoting a client’s community while writing ‘I’m part of the group,’ and does not mention disclosure.

The video says the comments are “trying to promote basically a community for one of my clients,” and every example ask is written in the first person as a member. It never says the author is paid to do this. That is the exact case the FTC’s Guides Concerning the Use of Endorsements cover, and it is also just a worse position to be in when a lead asks “wait, is this your group?” The fix is three words.

Lesson 8

Measure it, and what to change when it is not working

One sheet, one row per comment. Without it you will remember the two replies and forget the sixty comments, and you will not know which posts, which asks, or which hours are working.

The columns

Date. Post link. Post author. What the post was about, in five words. Which ask you used. Reply from the author (yes/no). Replies from anyone else (count). Connection requests received that day. Leads (someone asked for the free thing). Conversations (someone asked about the paid thing). Clients.

The weekly review

Every Friday, twenty minutes. Which posts got replies: were they from your thirty accounts or from search? Which asks got replies: was it the phrasing or the topic? What time of day? Then change one thing for the following week, not four.

Reading the numbers

  • Comments up, replies flat: the comments are not specific enough, or the posts are too old by the time you get there. Re-read lesson four; move to the first two hours.
  • Replies up, leads flat: the ask is not compelling, or it is not clearly free. Make the front end more concrete: a number of things, a named outcome.
  • Leads up, conversations flat: the front end is not leading anywhere, or you are not following up. Lesson seven, the next two weeks.
  • Conversations up, clients flat: this is a sales problem, not a comment problem, and it is out of scope here. The good news is that you have a pipeline.

Checked The source reports no numbers at all, so there is no benchmark to give you.

“My reply rates and the conversion rates are pretty good as well” is the whole of it. Two screenshots of replies are shown. We have no figure to pass on and will not invent one; the first month of your own sheet is the benchmark, and it will be better than any number from someone else’s niche.

Lesson 9

The whole thing on one page

Once

  1. Write the one-page offer document: buyer, five struggles, the free front end, the paid back end.
  2. Build the front end so it can be sent within a day of a yes.
  3. Find and follow the thirty accounts your buyers read. Turn on the bell.
  4. Save five or six post searches in your buyer’s words.
  5. Set up the sheet.

Daily, twenty minutes

  1. Notifications first: any post from the thirty under two hours old.
  2. Saved searches, past week, five posts with real discussion.
  3. Read the post and its comments. Draft with AI if you like; check every fact; rewrite the ask; post by hand.
  4. Two paragraphs of specific value, one free ask in the post’s terms, under 120 words, no link.
  5. Log it.
  6. Answer every reply within the hour. Connect, send the thing, no pitch.

Weekly

  1. Follow up once with everyone who took the free thing.
  2. Friday review. Change one thing.

Never

  1. Scrape posts or profiles.
  2. Let anything post for you.
  3. Use the same opening line or the same ask twice in a day.
  4. Put a paid offer in a comment, unless the post asked for a vendor.
  5. Promote something you are paid to promote without saying so.

Checked What the source promised and did not deliver.

The video ends by saying “complete SOPs of this process” and a step-by-step document are linked in the description. The description contains no link. This page is that document, with the parts that break LinkedIn’s rules taken out.

Written by

Alwaz Tahir, founder of ScaleWise VA, a Shopify operations and growth agency. The agency gets its own clients the way this course describes, among others, and the rules in it are the ones the agency works to.