# How to use AI for LinkedIn social selling
> Use AI to find LinkedIn commenters who already have the problem you solve, reply with a useful tip, and turn the conversation into a meeting.
- HTML: https://getkinlead.com/how-to-use-ai-for-linkedin-social-selling
- Markdown: https://getkinlead.com/how-to-use-ai-for-linkedin-social-selling.md

**Short answer:** Social selling with AI means finding people who are already talking about the problems you solve. You read their comments, reply with a useful tip, and only ask for a meeting after they engage.

## Stop cold pitching

Most LinkedIn outreach starts with a list of job titles and a message about your product. The buyer has no reason to care. They haven't said they have the problem, and they don't know you. A pitch in that moment is noise.

Social selling works the other way. You look for a person who is already describing the problem you solve. Maybe they commented on a competitor's post about a slow API, or asked a group how to fix a broken onboarding flow. You reply with something useful. The pitch comes later, if it comes at all.

On the first pass, ignore your offer. Look for a sentence you can respond to. You have to read the comments closely. Scanning for names won't tell you who has a live problem.

**Cold pitching**

- Sending the same 'I help teams grow' message to 50 people who have never heard of you.
- Focusing on your product features and your company history.
- Asking for a 15 minute call in the first message.

**Social selling**

- Replying to a post where someone says their API lags at 1,000 requests with a tip about cache clearing.
- Asking what broke before you mention your product.
- Sharing a public tip, then waiting for them to reply before sending a connection request.

## Find people already talking about the problem

You don't need to guess who is ready to buy. People leave clues in public. They ask for a tool recommendation, complain about a current vendor, or describe a workaround that stopped working. Your job is to find those comments before someone else does.

Pick 5 to 10 company pages or people who post for your buyers. Competitors work well because their followers are already using a product in your category. Industry analysts and newsletter writers work too. Open the comments on their last few posts and read. You are looking for a specific sentence. A general feeling about the post won't help.

Work in batches. Pick one competitor page, read the last five posts, and write down every comment that names a problem. Then move to the next page. If you try to read everything at once, you'll stop at the first interesting post and never build a list. Set a 20 minute timer. You want three to five live problems you can reply to today.

1. **Identify target profiles.** Choose competitors, industry newsletters, or people whose followers match your buyer. A 40-person agency selling to ecommerce brands might watch three competitor pages and two agency founders.
2. **Read the last 10 posts.** Open comments and look for questions, complaints, or workarounds. Skip praise comments like 'Great post' because they don't tell you anything.
3. **Copy the comment and the post topic.** Keep the exact words. You'll need them when you write the reply and the connection request.
4. **Sort by signal.** Put live problems at the top, tool questions next, and general comments at the bottom. Work the top of the list first.

| | Comment type | What it usually means | What to do |
| --- | --- | --- | --- |
| Asking for a recommendation | They are open to switching | Reply with one honest option and a question about their setup |
| Complaining about a current tool | They have a live problem | Ask what broke, then offer a fix that doesn't need your product |
| Sharing a workaround | They are solving it manually | Say what you'd check first, then ask if they want the short version |
| Tagging a colleague | They are comparing notes internally | Don't jump in. Save the post and watch for a follow-up. |

## Draft a reply that helps before it sells

Once you find a live problem, write a reply that would be useful even if you never mentioned your company. The reply should show you read the comment. It should give one concrete thing to check or ask one question that moves the conversation forward.

Use AI to draft options. Don't let it send blind. Paste the comment and the post topic. Tell the AI to write two sentences, no product mention, no calendar link, and a casual tone. Then edit it so it sounds like you. If the draft says 'I understand your frustration' or 'That sounds challenging', delete it. Those lines are filler.

If the comment is vague, ask a clarifying question instead of guessing. 'Is the lag on every request or only after a deploy?' That keeps you from writing a tip that misses. The reply should stand on its own. If the person never clicks your profile, they should still get value from the comment. That is the test. If your reply only makes sense as a setup for a pitch, rewrite it.

**Weak reply**

- I understand your frustration with API lag. That sounds challenging.
- We help companies fix API lag with a better caching layer. Want to see a demo?

**Strong reply**

- That usually happens when the cache isn't cleared after peak traffic. Are you clearing it on a schedule or only when the lag starts?
- If it's only when the lag starts, the first request after a spike will still be slow. Happy to share the check we use.

- **them:** I'm so tired of our API lagging every time we hit 1k requests. It's killing our onboarding flow.
- **draft:** That usually happens when the cache isn't cleared every 24 hours. Have you tried clearing it after peak traffic?
- **them:** I hadn't tried that. Let me check. Thanks for the tip.

```
Prompt: Here is a LinkedIn comment: [paste comment]. The post was about [topic]. My product helps [buyer] with [problem], but do not mention my product. Write two replies. Each reply should be two sentences, give one practical thing to check, and ask one question. Tone: peer to peer, no 'I understand', no 'sounds challenging', no calendar link.
```

## Send the connection request after they reply

Wait for a reply before you send a connection request. If they answer your comment, they know your name. If they don't, a request looks like every other cold invite. The reply is the signal that you can move to a private message.

Keep the request about the conversation. Mention the topic. Leave your product out of it. If you add a note, keep it short and write like a person. 'Enjoyed your point about the API lag. I'm connecting with people working on onboarding flows. Cheers.' That works because it references the shared topic and gives a reason for the request.

If they accept without replying to the comment, you can still send a message that references the post. But don't pretend you had a conversation that didn't happen. If they accept but don't reply, wait two days. Then send one message that references the post. Leave the invite out of it. 'Your comment about manual reporting stuck with me. Did you find a fix, or is it still a spreadsheet problem?' That starts a new conversation and gives them an easy answer.

1. **Wait for a reply.** Give it a day. If they answer, send the request. If they don't, leave it. You can keep engaging with their future comments.
2. **Write a short note.** Use the topic, not your pitch. 'Your comment about manual reporting stuck with me. I'm connecting with ops leads at agencies. Happy to keep swapping notes.'
3. **Skip the calendar link.** The invite is not the place for a demo ask. Save that for after they accept and you've sent one useful message.
4. **Log the lead.** Put the person, the post URL, and the comment in a simple sheet or doc. You'll want the context when they accept.

> If they don't reply to your comment, don't send the request. It will look like another cold invite.

## Write the first message after they accept

After they accept, send one message that continues the conversation. Don't start over with 'Thanks for connecting, I help companies with X.' That resets the trust you built. Reference the comment or the tip you shared.

Ask a small question that helps you qualify without an interrogation. 'Did clearing the cache fix the lag, or is it still showing up under peak load?' That question is easy to answer, and the answer tells you whether the problem is still live. Keep the first message short. Two sentences is enough. If you write a paragraph, it reads like a pitch even when it isn't. The goal is one reply. Save the discovery call for later.

If they don't answer the first message, send one follow-up with a new useful detail. Don't ask 'just bumping this'. If they ask a question you can't answer, say you'll check and come back. Then do it.

- **them:** Thanks for connecting.
- **you:** Thanks for accepting. Did clearing the cache fix the lag, or is it still showing up under peak load?
- **them:** It helped for a day, then came back when we hit 2k requests. So still looking.
- **you:** That points to the queue, not the cache. Are you using a single queue for onboarding and background jobs?
- **them:** Same queue. That's probably the issue.

> Send one useful message, then ask one question. Don't stack three questions in the first reply.

## Doing this with Kinlead

Kinlead does the watching and the first draft. The [Steal Customers agent](/features/steal-customers) watches competitor company pages and optional founder or employee profiles, reads comments on their recent posts, and turns commenters who look like buyers into leads. Each lead keeps the post URL, what the post was about, and what they commented.

Lead search uses LinkedIn's standard search, so you don't need a Sales Navigator seat. AI can write connection notes under 200 characters that open with why that person became a lead. Kinlead writes messages and follow-ups from your campaign goal, tone, and an optional sales playbook.

Kinlead sends from your own LinkedIn account through Unipile. It sends at a human pace, at most one action every 5 minutes, inside daily invite and message limits and a send window you set. Every campaign's replies land in one inbox. Kinlead classifies them by intent and puts interested replies first. You can hand the conversation to yourself at the first reply, let the AI reply until there is real interest, or let it reply until a meeting is booked with your scheduling link. You can read and edit queued campaign messages on the Actions page before they send, and pause any campaign.

## Related

- [AI LinkedIn outreach](https://getkinlead.com/features/ai-linkedin-outreach.md)
- [Lead groups and scoring](https://getkinlead.com/features/lead-groups-and-scoring.md)
- [Pricing plans](https://getkinlead.com/pricing.md)

## Common questions

**Will using AI for LinkedIn get my account banned?**

LinkedIn does not publish exact limits, and any automation carries risk. Keep daily actions low, use delays that look like a person typing, and stop when LinkedIn pushes back.

**How many connection requests can I send per day?**

LinkedIn doesn't share a fixed number. Set a limit you can keep for weeks, watch your acceptance rate, and lower it if people stop accepting.

**Do I need Sales Navigator to find these comments?**

No. Watching comments on competitor pages and industry posts works from a standard LinkedIn account.

**What if they reply to my comment but don't accept the connection request?**

Keep engaging in public. Don't send a second request. If they comment again, reply with another useful point and wait for them to accept or ask a question.

**How do I handle replies once the AI has started the conversation?**

Decide before you send. For most teams, hand off at the first reply so you can read the context and answer as yourself. If you let the AI continue, check the thread daily and step in when the person asks about price, timing, or a meeting.

[Create a Kinlead account](https://getkinlead.com/signup) or [book a call](https://getkinlead.com/demo) to talk through pricing and setup.
