// PLAYBOOK
LiveHow to Score LinkedIn or Sales Navigator Leads Against Your ICP and Draft Personalized Outreach
// WHEN TO USE
Problem
LinkedIn or Sales Navigator search returns hundreds of profiles, and most of them are not a fit.
Teams handle this in two expensive ways.
Some message everyone, which burns limited sending capacity, hurts reply rate, and trains them to blast more.
Others pay to enrich and AI-score every single row in a tool like Clay, spending credits on leads that were never going to qualify. Either way you pay full price for the whole list before you know who is worth it.
// WHAT YOU GET
Outcome
A funnel that spends effort only where it pays off. It collects clean, structured profiles from Periodix Search (name, title, company, location, and summary for every result), applies free rule-based filters to drop the obvious non-fits, then lets AI score only the survivors into validated and rejected, and drafts a personalized connection note and first message for the validated ones alone.
You do not blast everyone, and you do not pay to AI-score the whole list.
Validated leads land ready to send; rejected ones are logged with a reason to sharpen your ICP.
Waterfall outreach is usually expensive; run cheap-to-expensive, it is not.
// RESULTS
Proof
"We used to have SDRs do all of this by hand: pulling leads, scoring them against our ICP, and sending the same cold template. This workflow replaced that end to end, and now every first message is personalized, not a template. The output is as good as our best rep, and it runs every day without dropping a lead."
// BUILDING BLOCKS
Periodix Actions used
LinkedIn & Sales Navigator Search
Run a LinkedIn or Sales Navigator search by URL and get clean, structured rows back. Works for People, Companies, Posts, and Jobs on both Classic and Sales Navigator. Paste a search URL, and Periodix Actions returns the matching results with automatic rate limiting. This is the core of any LinkedIn lead generation or list-building workflow.
Create Sequence
Create a multi-step LinkedIn outreach sequence: a series of message templates with a delay before each step, an optional connection request before the first message, and a sending schedule by day and hour in UTC. Templates support {{variable}} placeholders with fallbacks, such as {{first_name|there}}, filled from each recipient's data when you enroll them. A recipient from Search People gives you {{first_name}}, {{headline}}, and the other fields of that record. The first step's delay counts from the moment the recipient accepts the connection request; each later step counts from the previous one. Nothing is sent until you add people with Enroll Recipients. Available in MCP and REST. Coming soon to the n8n node.
Enroll Recipients
Start a sequence for 1 to 500 recipients in one call. Each entry names the sending profile, the recipient, and the variables its message templates need. Recipients come from Search People, Get Profile, or Get Connections: use their provider_id as the recipient. Every {{variable}} in the sequence without a fallback must be supplied for each recipient. Nothing is sent during the call. Messages go out on the sequence schedule and within the sending limits of each profile, so a large list is delivered gradually. You choose which profile sends to which recipient, so you can split a list across several accounts. Available in MCP and REST. Coming soon to the n8n node.
// HOW IT WORKS
Flow
// RUN IT
Implementations
This workflow runs on Periodix Actions, so you need an API key and a connected LinkedIn profile first.
- Start your free trial (3 days, full access) and connect a LinkedIn profile: https://actions.periodix.net/dashboard/register
- Follow the step-by-step recipe: https://docs.actions.periodix.net/docs/n8n-recipe-sales-navigator-icp-scoring
- Import the ready-made template json into n8n: https://n8n.io/workflows/16373-score-linkedin-icp-leads-and-draft-outreach-with-claude-notion-and-sheets/
What you need: the verified Periodix Actions node, an Anthropic credential, a Notion database, and a Google Sheet with Validated and Rejected tabs. Sales Navigator URLs need a Sales Navigator seat on the connected profile; standard LinkedIn search URLs work with any account.
Setup: create a Notion database with Sales Nav URL, Status (Pending, Done), Campaign, and Processed At; pick your connected profile; connect Notion and Google Sheets; add your Anthropic credential and your product and ICP to both prompts; add search URLs to Notion as Pending.
Every hour the workflow runs each pending search, scores every lead 1 to 10 against your ICP, writes a connection note and first message for leads scoring 5 or higher, and marks the Notion row as Done.
// OUTPUT SHAPE
Example JSON output
{
"linkedin_url": "https://www.linkedin.com/in/jane-doe",
"full_name": "Jane Doe",
"headline": "VP Sales at Acme",
"company": "Acme",
"location": "San Francisco",
"icp_score": 9,
"reason": "VP-level at a 200-person SaaS, owns the buying decision for our category.",
"connection_note": "Hi Jane, saw you lead sales at Acme ...",
"message": "Following up with a quick thought on ...",
"status": "validated"
}// CONNECTS WITH
Integrations used
OpenAI
Google Gemini
Google Sheets
Notion
Claude (Anthropic)// FITS INTO
Use cases
// QUESTIONS
FAQ
- Why score leads before outreach instead of after?
- Your sending capacity is limited and personalization is only worth it on real prospects. Scoring first concentrates your effort on leads that can convert.
- How does AI scoring read a profile?
- It reads the whole profile, headline, summary, role, and company, and returns a score from 1 to 10 with a short reason, given your ICP description.
- Why score leads before outreach instead of after?
- Your sending capacity is limited and personalization is only worth it on real prospects. Scoring first concentrates your effort on leads that can convert.
- How does AI scoring read a profile?| What happens to rejected leads?
- They are logged with their score and reason. That rejected list is a mirror of your ICP and is how you refine your targeting over time.
- Do I need Notion and Google Sheets?
- The template uses Notion for input URLs and Sheets for output, but you can swap either for your own CRM or database.
- How is this different from Clay?
- Clay charges credits to enrich and score every row, so you pay across the whole list. This playbook drops obvious non-fits with free rules first, scores only the survivors with a cheap AI call, and drafts messages only for validated leads, so you spend on scoring and personalization where it matters, not on people who were never a fit.
- What does Periodix Search return, and how does the AI use it?
- Periodix Search returns clean, structured profiles for every result: name, headline, role, company, location, summary, and profile URL. The AI reads that structured data and scores each lead against your ICP, so the accuracy of the scoring comes directly from the rich, consistent fields the search returns, no manual cleanup needed.
- Can I use OpenAI or Gemini instead of Claude?
- Yes. The scoring and drafting steps are standard model calls, so you can run them on Claude, OpenAI GPT, or Google Gemini. The published template uses Claude, but swapping the model does not change the rest of the workflow.
// EXPLORE MORE
Related playbooks
Turn hiring signals into outbound leads
When a company posts a job for a role you sell around, it is telling you it has a new priority, a fresh budget, and a pain. That is one of the strongest buying signals in B2B. The trouble is that job posts are scattered across LinkedIn, most are from recruiting agencies and staffing firms you do not want, and the person who posted is a recruiter, not the decision-maker you need to reach. Working this by hand does not scale, and by the time you get to a post it is often stale.
Mine LinkedIn post engagement for intent signals
The people who like and comment on posts in your niche are already raising their hand, and that intent is spread across dozens of posts and buried in the feed. To work it by hand you first have to find the right posts, then open each one, scroll every reaction, read every comment, and copy profiles. It does not scale, so the warmest signals go to waste and teams keep buying cold lists instead.
// FROM THE BLOG
Related articles
Best LinkedIn MCP Servers in 2026: Which One Fits Your Job
Deep vs Wide LinkedIn Enrichment Without Clay: Get Profile for Research, Sales Navigator Search for Scale
AI Lead Scoring: How to Qualify LinkedIn Leads by ICP Before You Reach Out
Cold vs ICP-Scored vs Signal-Based LinkedIn Outreach: 10-20% vs 33% vs 81% Acceptance
// GET STARTED
Start building with Periodix Actions
Create an account, get an API key, and run this playbook end-to-end.
Create your account