AI for Outbound Webinar March 19th
A recorded live webinar where you watch a full AI outbound pipeline get built: source leads with Apify, clean them with Claude, research personal hooks with Perplexity, and prepare LinkedIn outreach, for about $60 a month in tools.
Pour qui: Business owners and operators who know they should do outbound but find sourcing, cleaning and researching leads by hand too slow.
À retenir
Outbound is four steps
Every outbound funnel has four parts: lead sourcing (finding the people you want to talk with), data cleaning (the lists come back messy), research (checking each person is worth contacting) and contact. Each one takes a lot of time by hand, and email-only outreach at volume can burn your domain and get you blacklisted. The webinar speeds up each step with AI.
Automate everything except the message
The contact step stays with you. If you automate a wrong message sent to the wrong person on the wrong channel, you just scale something that does not work. Automate finding, cleaning and researching first. Only automate the contact itself once you know your email, LinkedIn or other channel is getting results, and then only to add volume.
Source leads with the Apify Leads Finder
Instead of scrolling lists or Sales Navigator, you give Apify's Leads Finder actor a job title, location, company size or keywords and it returns matching people. In the demo, searching for CEOs in Mexico and asking for 100 of them took seconds. Leads also live on Google Maps (local businesses), Instagram, YouTube comments and Reddit, where people describe their problems.
Let AI clean the dirty data
Raw lead lists always have gaps: missing emails, missing last names, no phone numbers. Instead of working through it in Excel, download the CSV to a folder, point Claude Cowork at that folder and ask it to clean the data and build a dashboard. It removed duplicates and blanks and returned a clean CSV of leads with emails in seconds, plus an interactive dashboard.
Research people to find a hook
Before contacting someone, run deep research on them with Claude or Perplexity to find something about their recent career or life you can mention, or intent signals such as engaging with your competitors. One client's campaign looks for people who recently appeared on a podcast and invites them to another. Deep research on one person takes 3 to 5 minutes, so save it for people you really want to reach.
Research the whole list in parallel
For research at scale, I asked Claude Code (connected to my Apify account) to use Perplexity on every lead and look for recent news that could serve as a hook. It wrote a parallel script, handling many leads at once instead of one at a time, and finished 13 people in 14.6 seconds with zero errors. The results work as personalization hooks or as a filter to drop leads.
The math and the cost
Apify at $39 a month plus Claude at $20 gets you this pipeline for about $60. If you send about 100 LinkedIn connections a day, roughly 20 percent accept and maybe 5 percent reply, which is already five potential business conversations. Because you also have emails, you can add a cold email tool such as Instantly and reach people on both channels.
Outils présentés
- ApifySourcing leads with the Leads Finder actor (job title, location, company size, keywords), plus scraping LinkedIn post reactions, comments and profiles
- Claude (Cowork and Claude Code)Cleaning the raw lead CSV, building a lead dashboard, running deep research on a person, writing the parallel research script and creating the Google Sheet
- Claude in Chrome extensionGoing through the lead sheet and sending LinkedIn connection requests from the browser, without the LinkedIn API
- PerplexityResearching every lead and their company for recent news to use as a personalization hook
- Google SheetsHolding the cleaned leads and the hook summaries Claude produced, connected to Claude through the built-in connectors
- InstantlyRecommended cold email system to reach the same leads by email alongside LinkedIn
- ExaMentioned as another good option for automating lead research
- PostHogBonus: website analytics that Claude analyzed to find rage clicks and bugs on my site
Questions du public
Does LinkedIn have an API you can connect Claude Code to?
It does, but it is extremely expensive, so I work around it. When I reached out to leads in the demo I was not connecting to LinkedIn through an API. I used Claude to interact with my own LinkedIn account, with an agent clicking buttons to send connection requests. You are not connecting your LinkedIn account to anything, and you avoid paying thousands of euros for API access.
Is there a simpler way to connect Claude to Google Sheets than an MCP server?
Yes. In the Claude chat window, click the plus symbol, open connectors, and pick the ones you need, such as Drive search, Canva, Gmail, Google Calendar and Google Sheets. Once the Claude app is connected, Claude Code also has access to those connections, so you do not have to deal with MCP servers. Even without Claude Code, a connected Claude chat can create the sheets for you.
Will using the Claude Chrome extension on LinkedIn get my account restricted?
I have never heard of anyone hitting restrictions on LinkedIn from using the Claude Chrome extension, which surprised me, and that is why I allow myself to use it much more. I still recommend keeping the actual outreach in your hands until you know your message works.
Can Apify get leads from LinkedIn without connecting my account?
Yes. Apify can scrape the reactions on any LinkedIn post, the comments, and people's profiles without you ever connecting your LinkedIn account, and it is very cheap. That is one of the main reasons I like Apify for anything related to LinkedIn.
À regarder ensuite
9:46How to Get Local Business Leads with AIConnect an Apify scraper to Claude and pull your first lead list with contact details.
6:44Get Your Money's Worth from Perplexity (2026 Guide)Run deep research in Perplexity to find a personal hook for each prospect.
7:38Use Cowork Subagents to 10x Your ProductivityRun several subagents in parallel so a whole list gets researched at once.
