Coconote
AI notes
AI voice & video notes
Try for free
🤖
AI Strategies for Cold Email Personalization
May 13, 2025
Lecture on AI Personalization System for Cold Emails
Introduction
Topic
: AI personalization for cold emails
Goal
: Create a cost-effective system for better reply rates
Offer
: Integration available through link in description
AI Personalization Systems
Basic Principle
Two Components
:
Personalization Data
: Information from websites, LinkedIn, etc.
LLM Utilization
: Use language models (e.g., Chat GPT) to create personalized messages
Approaches
Integrated Approach
:
Combines data collection and personalization in a single API
Pros
: Fast, already available information
Cons
: High cost for web searches
Separated Approach
:
Individual steps for data collection and personalization
Solution Highlighted
: Tavi & Crawl for AI
Tool Comparison
Tavi
Functionality
: Gathers search data in a format suitable for LLMs
Use Case
: Query specific data (e.g., case studies)
Crawl for AI
Features
:
LLM-friendly web scraper
Open source, can be hosted to save costs
Cost Comparison
:
Cheaper than Clay.com ($800/month for 50k credits)
Crawl for AI with GPT-4.1 nano costs $22 for 50k leads
Clay.com
Pricing
: $800/month for 50k credits
Comparison
: 4x more expensive than Crawl for AI solution
Workflow Structure
Steps
Lead List Check
Look for personalization opportunities
Extract domain and build Crawl for AI configuration
Crawl for AI Execution
Scrape data from 10-15 pages
Pass information to OpenAI for email personalization
Update Google Sheet with results
Tools & Components
URL Pattern Finder
:
Prioritizes URLs with specific patterns
Keyword Scorer
:
Assigns weight to pages with keywords (e.g., testimonials, case studies)
Allows for dynamic and relevant personalization
Output
Array of markdown data
Personalized email lines created by OpenAI
Conclusion
System
: Inexpensive and effective for scaled email personalization
Availability
: Contact to integrate into business
Closing
Next Steps
: Reach out for integration
Goodbye Message
: Video end
📄
Full transcript