AI for LinkedIn Webinar March 6th
This webinar shows you three ways to use AI on LinkedIn: handle your inbox with the Claude Chrome extension, pull leads from other people's posts with Apify's cookieless scrapers and Claude Code, and sync every post you have published to Notion through Unipile so you can see what performs.
Who it's for: Business owners, founders and operators who use LinkedIn for leads or content and want to automate the tedious parts without putting their account at unnecessary risk.
Key takeaways
Claude Chrome handles low-volume LinkedIn tasks
The Claude extension for Chrome works best for slow, low-volume jobs on platforms with no API or strict automation rules. I ask it to extract my last ten LinkedIn messages and log them in a spreadsheet, and it works across both tabs. I also used it to find the LinkedIn profiles for 125 webinar sign-ups from their names and companies. It is slow by design, but you can run several at once.
Teach Claude once, then schedule it as a task
You don't need a special skill for LinkedIn. The Teach Claude button records you doing the job, Claude analyzes the video, and it turns what it saw into a task. Type a slash in the extension to schedule a task, then give it the prompt and the page to start from. I have one that likes comments on my profile at 10 AM every morning.
Keep your LinkedIn risk low
Every automation carries some risk, because LinkedIn's terms are strict. To lower it, ask Claude Chrome or Claude Code to wait a random amount of time between actions when working through a long list. I only automate things I could do by hand at a normal pace, plus data collection. I do not do automated LinkedIn messaging. Tools that connect your own account, such as Unipile and PhantomBuster, carry risk.
Cookieless scrapers mean zero account risk
Apify is a library of scrapers. The ones marked "no cookies" never connect your LinkedIn account; someone else's account does the scraping, so your risk is zero. I combine separate scrapers for post comments, reactions, reposts and profiles. That shows you who engaged with a post, and you can mention that engagement when you write to them.
Build a "parasite" lead system from other people's posts
Start with one post: scrape its comments, then have Claude Code scrape each commenter's profile and pick out your persona, for example CEOs of companies in Europe. Then expand it: pull every post a creator published in the last 10 to 30 days, scrape all the engagement, add company data, and flag the people who interact often. That gives you several intent signals per lead.
Turn your own posts into a content database
Unipile ($55 per month) connects your LinkedIn account through an API key. I ask Claude Code to sync all my LinkedIn posts to a Notion database with impressions, reactions, engagement rate, image and link, then to analyze which hooks, themes and visuals perform best. To go further, save the metrics every day and track how each post's numbers change.
Two tools run the whole lead machine
For a full LinkedIn lead finder you need only Apify and Claude Code (or Antigravity). Use Apify to search LinkedIn for lead-magnet posts that still get high engagement, scrape their reactions, comments and shares, then scrape those people's profiles and companies. Claude Code qualifies them. Collect the best lead magnets in a database and the same flow shows you what content works in your niche.
Tools shown
- Claude Chrome extensionLow-volume browser tasks: logging LinkedIn messages to a spreadsheet, finding LinkedIn profiles for a list of names, and running scheduled tasks such as liking comments
- Claude CodeSending LinkedIn messages through Unipile, running Apify scrapers, qualifying leads and analyzing post performance
- ApifyCookieless LinkedIn scrapers for post comments, reactions, reposts, profiles and search, without connecting your account
- UnipileAPI connector for your LinkedIn account, used to send messages and sync all your posts with their metrics
- NotionContent management database that holds every LinkedIn post and its metrics
- PhantomBusterAutomatic LinkedIn actions that run when something happens (accepted requests, likes, follow-backs) and sending messages from a database
- AntigravityGoogle's Claude Code alternative, recommended if you are on a $20 per month budget
- Playwright MCPGiving Claude Code browser access for the profile viewer scraper
- Nano Banana ProCreating visuals for LinkedIn posts in a content machine
Questions from the audience
Does Claude need a special skill to understand LinkedIn?
No, it doesn't need a skill. In Claude Chrome you can use the Teach Claude button at the bottom right. It records your screen while you do the task yourself, clicking through the buttons. Claude analyzes that video and turns it into a task, which I describe as the equivalent of a skill for this version of Claude.
Is there a risk of getting punished by LinkedIn for using automation?
There is always some risk, and LinkedIn is always a touchy subject, but it is hard for them to track. You can ask Claude Chrome or Claude Code to wait a random amount of time between actions, for example when working through 80 or 100 messages. I only automate what I could do by hand at a normal pace, plus data collection. I don't do automated LinkedIn messaging. Cookieless Apify scrapers take the risk to zero for scraping.
Once I have the people who engaged or viewed my profile, how do I message them?
You can put them in a spreadsheet and hand it back to Claude Code. The simpler route is PhantomBuster: connect the database to it and it sends the messages for you. If you want messages sent automatically, without having to push a button yourself, I recommend PhantomBuster. First decide your exact workflow, then pick the tools that fit it.
How much does the lead magnet flow cost?
About $8 to $10 per 1,000 leads. You pay around $3 per 1,000 profiles on Apify, plus about $3 to $5 for the reactions and comments. These leads have shown intent: they engaged with something someone else is selling, so this is not random scraping.
