How to Build an AI-Powered LinkedIn Message Writer

I built this system to generate personalized linkedin messages using ai analysis of prospect profiles. This ai linkedin message writer personalized approach scales personalized outreach without losing the human touch.

Outreach that feels automated gets ignored. Outreach that feels personal gets replies. This delivers personal at scale.

What You Need Before Starting

• Python 3.8+ with requests
• Anthropic API key
• Email sending service (SendGrid or SMTP)
• Prospect database or CRM export

Step 1: Load Prospect Data

import csv

def load_prospects(path):
    prospects = []
    with open(path) as f:
        for row in csv.DictReader(f):
            prospects.append({
                "email": row["email"], "name": row["name"],
                "company": row["company"], "title": row.get("title", ""),
            })
    return prospects

Step 2: Research Each Prospect

import anthropic
from dotenv import load_dotenv

load_dotenv()
client = anthropic.Anthropic()

def research(prospect):
    message = client.messages.create(
        model="claude-sonnet-4-20250514",
        max_tokens=512,
        system="Research this prospect. Find personalization angles and likely problems.",
        messages=[{"role": "user", "content": f"{prospect['name']} at {prospect['company']} ({prospect['title']}). Find talking points."}]
    )
    return message.content[0].text

Step 3: Generate Personalized Outreach

def write_message(prospect, research_data, msg_type):
    message = client.messages.create(
        model="claude-sonnet-4-20250514",
        max_tokens=512,
        system="Write personalized outreach. Under 100 words. One clear ask. No generic flattery.",
        messages=[{"role": "user", "content": f"To: {prospect['name']} at {prospect['company']}. Research: {research_data}. Type: {msg_type}."}]
    )
    return message.content[0].text

Step 4: Schedule Follow-Ups

import sqlite3
from datetime import datetime, timedelta

db = sqlite3.connect("outreach.db")
SCHEDULE = [3, 7, 14]  # days

def schedule_followups(email, initial_msg):
    for i, days in enumerate(SCHEDULE):
        send_at = datetime.now() + timedelta(days=days)
        db.execute("INSERT INTO followup_queue VALUES (?,?,?,?)",
            (email, f"followup_{i+1}", send_at.isoformat(), "pending"))
    db.commit()

What to Build Next

Add reply detection. When a prospect replies, pause their sequence and alert you to respond personally.

Related Reading

• AI in Paid Advertising: The Complete Overview - operational AI systems that run themselves
• Why AI Operations, Not AI Tools - operational AI systems that run themselves
• AI for Creative Strategy and Testing - practical guide for business operators building AI systems