02 — LinkedIn Outreach Automation with AI Personalisation

Part of the LifeSciForge portfolio — 10 AI projects targeting 50 LPA roles in Life Sciences

An n8n workflow that automates LinkedIn connection requests for pharma and life sciences professionals. It fetches live LinkedIn profile data, detects the prospect's functional area (Medical Affairs, HEOR, Market Access, etc.), generates a personalised 100–120 character connection note using Claude AI, validates message quality, checks connection status, sends the request with human-like random timing, and logs everything to Google Sheets.

🎯 What This Workflow Does

Triggers Monday–Friday at 11am automatically
Reads unsent contacts from Google Sheets
Applies a random daily limit of 8–12 connections (human-like behaviour)
Fetches live LinkedIn profile data via Browserflow
Detects functional area from job title (7 pharma-specific categories)
Generates a personalised AI connection note using Claude API (100–120 chars)
Validates message quality across 5 checks before sending
Checks connection status — skips if already connected or pending
Sends connection WITH personalised message (falls back to blank if unsupported)
Logs every outcome to Google Sheets: YES + MESSAGE, YES (BLANK), or SKIPPED
Waits randomly between contacts (10–25 min) to mimic human behaviour

🏗️ Workflow Architecture

Schedule Trigger (Mon-Fri, 11am)
        │
        ▼
Generate Random Time (1 min startup delay)
        │
        ▼
Wait Until Start Time
        │
        ▼
Get Unsent Connections (Google Sheets)
        │
        ▼
Apply Random Limit (8-12 per day)
        │
        ▼
Process One by One (SplitInBatches)
        │
        ▼
Get LinkedIn Profile Data (Browserflow)
        │
        ▼
Wait After Profile (2 min)
        │
        ▼
Identify Functional Area
[medical affairs / market access / clinical / regulatory / commercial / bd / heor]
        │
        ▼
Generate AI Message (Claude Sonnet)
        │
        ▼
Validate Message Quality
[length 100-120 / has AI reference / first person / starts Hi Name, / ends with ?]
        │
   ┌────┴────┐
PASSED     FAILED
   │           │
   ▼           ▼
Check       Skip to
Connection  Next Person
Status
   │
   ├── Already connected/pending → Update Sheet (SKIPPED) → Wait → Next
   │
   ▼
Calculate Pre-Send Wait (15-35 min random)
        │
        ▼
Wait Before Send
        │
        ▼
Try Send Connection WITH Message (Browserflow)
        │
        ▼
Wait After Send (2 min)
        │
        ▼
Check Send Success
        │
   ┌────┴────┐
SUCCESS    FAILED
   │           │
   ▼           ▼
Update     Send Blank
Sheet      Connection
(YES+MSG)  Request
   │           │
   ▼           ▼
Calculate  Update Sheet
Next Wait  (YES BLANK)
   │           │
   └─────┬─────┘
         ▼
   Wait Next Person
   (10-25 min random)
         │
         ▼
   Back to: Process One by One

🧩 Nodes Used

Node	Type	Purpose
Schedule Trigger	Trigger	Runs Mon–Fri at 11am
Generate Random Time	Code	Startup delay logic
Wait Until Start Time	Wait	Pause before processing
Get Unsent Connections	Google Sheets	Read contacts where status is blank
Apply Random Limit	Code	Filter to 8–12 unsent contacts per day
Process One by One	SplitInBatches	Process contacts sequentially
Get LinkedIn Profile Data	Browserflow	Fetch live profile name, title, company
Wait After Profile	Wait	2 min pause after profile fetch
Identify Functional Area	Code	Map job title to pharma function
Generate AI Message	LangChain LLM Chain	Prompt Claude to write connection note
Claude AI Model	Anthropic (Claude)	claude-sonnet-4-20250514, temp 0.7
Validate Message Quality	Code	5-check quality gate
Message Quality Check	IF	Route pass/fail
Check Connection Status	Browserflow	Check is_connection and is_pending
Wait After Check	Wait	1 min pause
Can Send Request	IF	Route send/skip
Calculate Pre-Send Wait	Code	Random 15–35 min wait
Wait Before Send	Wait	Human behaviour simulation
Try Send Connection WITH Message	Browserflow	Send with AI note
Wait After Send Attempt	Wait	2 min pause
Check Send Success	Code	Detect send success or failure
Message Sent Successfully	IF	Route success/fallback
Update Sheet – Sent with Message	Google Sheets	Log YES + MESSAGE
Send Blank Connection Request	Browserflow	Fallback — send without note
Wait After Blank Send	Wait	2 min pause
Update Sheet – Sent Blank	Google Sheets	Log YES (BLANK)
Calculate Next Person Wait	Code	Random 10–25 min inter-contact delay
Wait Next Person	Wait	Inter-contact pause
Update Sheet – Skip	Google Sheets	Log SKIPPED
Calculate Skip Wait	Code	Random 5–15 min skip delay
Wait Skip	Wait	Skip pause before next contact

📊 Google Sheets Setup

Create a Google Sheet with the following columns in Sheet1:

Column	Description	Example
`First Name`	Prospect first name	Rajesh
`Last Name`	Prospect last name	Sharma
`Person Linkedin Url`	Full LinkedIn profile URL	https://linkedin.com/in/rajesh-sharma
`Title`	Job title (used for fallback area detection)	Head of Market Access
`Company`	Company name	Novartis
`LinkedIn Connection Sent`	Status — leave blank for new contacts	YES + MESSAGE
`Connection Request Date`	Auto-filled by workflow	2026-03-15
`Message Sent`	Auto-filled — exact message sent	Hi Rajesh, we have developed...
`Connection Accepted Date`	Fill manually when they accept	2026-03-18

Important:

Leave LinkedIn Connection Sent blank for contacts you want to send to
The workflow filters to blank rows only — it will not re-send to contacts already processed
Person Linkedin Url is the unique key used for matching and updating rows

⚙️ Setup Instructions

Step 1 — Prerequisites

Before importing this workflow, ensure you have:

n8n running (local, cloud, or self-hosted)
A Browserflow account with LinkedIn connected
An Anthropic API key for Claude
A Google account with Sheets API access enabled in n8n

Step 2 — Import the workflow

Open n8n → click "Workflows" in the left sidebar
Click "Import from File"
Upload linkedin-outreach-ai-personalisation.json
Click "Import"

Step 3 — Set up credentials

Add the following credentials in n8n (Settings → Credentials → Add Credential):

Credential	Type	Where to get it
Google Sheets account	Google Sheets OAuth2	Google Cloud Console
Browserflow account	Browserflow API	browserflow.app → Settings → API Key
Anthropic (Claude) account	Anthropic API	console.anthropic.com → API Keys

Step 4 — Replace placeholders

Search for and replace these values in the workflow after importing:

Placeholder	Replace with
`YOUR_GOOGLE_SHEET_ID`	Your Sheet ID from the URL
`YOUR_GOOGLE_SHEETS_CREDENTIAL_ID`	ID from n8n credentials
`YOUR_BROWSERFLOW_CREDENTIAL_ID`	ID from n8n credentials
`YOUR_ANTHROPIC_CREDENTIAL_ID`	ID from n8n credentials
`YOUR_COMPANY`	Your company name (in Generate AI Message node prompt)
`YOUR_PRODUCT`	Your product name (in Generate AI Message node prompt)
`YOUR_WEBHOOK_ID_1` through `_8`	Any unique strings (e.g. `wait-start-001`)
`YOUR_N8N_INSTANCE_ID`	Your n8n instance ID
`YOUR_WORKFLOW_ID`	Auto-assigned after import

Step 5 — Add contacts to Google Sheets

Add your prospect list to the Google Sheet. Required columns: First Name, Last Name, Person Linkedin Url, Title, Company. Leave LinkedIn Connection Sent blank.

Step 6 — Test the workflow

Set the Schedule Trigger to Manual trigger temporarily
Run the workflow with 1–2 test contacts
Verify the AI message is generated and logged correctly
Check that Google Sheet updates with correct status

Step 7 — Activate

Switch the trigger back to Schedule and toggle the workflow to Active.

🔒 Safety & LinkedIn Compliance

This workflow is built with multiple safety features to reduce risk of LinkedIn account restriction:

Daily limit: 8–12 connections per day (well below LinkedIn's ~100/day limit)
Weekday only: Monday–Friday only — no weekend activity
Random timing: All waits are randomised to avoid predictable patterns
Duplicate check: Checks connection status before every send
Message validation: Validates quality before sending — no spammy messages
Sequential processing: One contact at a time — never bulk sends

⚠️ Disclaimer: This workflow uses LinkedIn automation via Browserflow. Use responsibly and within LinkedIn's Terms of Service. The author takes no responsibility for account restrictions.

🧠 AI Prompt Logic

The workflow uses Claude Sonnet to generate personalised connection notes. The prompt:

Identifies the recipient's pharma functional area from their job title
Selects function-specific pain points and benefits
Generates a 100–120 character message (strict limit for highest acceptance rates)
Always starts with Hi {FirstName}, and ends with a question (CTA)
References AI product capability relevant to their specific role

Supported functional areas: Medical Affairs · Market Access · Clinical Development · Regulatory · Commercial · Business Development · HEOR

📈 Expected Outcomes

Based on production usage:

Connection acceptance rate: 28–35% (vs ~18% industry average for cold outreach)
Daily volume: 8–12 connections per run
Monthly volume: ~200–240 new connections per month
Time saved: ~2–3 hours per week vs manual outreach

🛠️ Customisation Tips

Change daily limit: Edit minConnections and maxConnections in the Apply Random Limit node
Change trigger time: Edit triggerAtHour in the Schedule Trigger node
Add new functional areas: Add new entries to departmentMap in Identify Functional Area
Adjust message tone: Edit the prompt in Generate AI Message node
Change message length: Adjust the length check in Validate Message Quality

📁 Repository Structure

LinkedIn_Outreach/
├── linkedin-outreach-ai-personalisation.json   # n8n workflow (sanitised)
├── README.md                                    # This file
└── Workflow_Documentation.docx                  # Full setup guide

🔗 Related Projects

01 — Email Outreach Automation
03 — Pharma CI Research Agent (coming soon)

👤 Author

Pranjal Das — CSO, Pienomial · Bengaluru, India

Part of the LifeSciForge portfolio — 10 AI projects built to demonstrate production-grade automation for life sciences roles targeting 50 LPA.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

02 — LinkedIn Outreach Automation with AI Personalisation

🎯 What This Workflow Does

🏗️ Workflow Architecture

🧩 Nodes Used

📊 Google Sheets Setup

⚙️ Setup Instructions

Step 1 — Prerequisites

Step 2 — Import the workflow

Step 3 — Set up credentials

Step 4 — Replace placeholders

Step 5 — Add contacts to Google Sheets

Step 6 — Test the workflow

Step 7 — Activate

🔒 Safety & LinkedIn Compliance

🧠 AI Prompt Logic

📈 Expected Outcomes

🛠️ Customisation Tips

📁 Repository Structure

🔗 Related Projects

👤 Author

About

Uh oh!

Releases

Packages

Uh oh!

Contributors

Uh oh!

Name		Name	Last commit message	Last commit date
Latest commit History 4 Commits
README.md		README.md
Workflow_Documentation_LinkedIn_Outreach.docx		Workflow_Documentation_LinkedIn_Outreach.docx
linkedin-outreach-ai-personalisation.json		linkedin-outreach-ai-personalisation.json

Folders and files

Latest commit

History

Repository files navigation

02 — LinkedIn Outreach Automation with AI Personalisation

🎯 What This Workflow Does

🏗️ Workflow Architecture

🧩 Nodes Used

📊 Google Sheets Setup

⚙️ Setup Instructions

Step 1 — Prerequisites

Step 2 — Import the workflow

Step 3 — Set up credentials

Step 4 — Replace placeholders

Step 5 — Add contacts to Google Sheets

Step 6 — Test the workflow

Step 7 — Activate

🔒 Safety & LinkedIn Compliance

🧠 AI Prompt Logic

📈 Expected Outcomes

🛠️ Customisation Tips

📁 Repository Structure

🔗 Related Projects

👤 Author

About

Topics

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Packages