Build a Cold Email Engine That Writes Hyper-Personalized Emails with AI

Every B2B founder knows cold email works. But writing 1000 personalized emails? That takes weeks. What if an AI could research each prospect and write a unique email -- not just "Hi {name}" but genuinely personalized based on their company, role, and recent activity? The Problem Generic cold emails get 2% reply rates. Personalized ones get 15-25%. But manual personalization at scale is impossible. The Solution: AI Research + Personalization Pipeline Step 1: Prospect List Start with a CSV: name, company, LinkedIn URL, website. Step 2: AI Research For each prospect, the agent scrapes their company website with Firecrawl, reads their LinkedIn summary, and identifies what they do, recent news, tech stack, and pain points. Step 3: Personalized Email Claude writes a unique opener for each prospect based on the research. Not template-y. Not generic. Actually relevant. Step 4: Send + Track Send via Resend with custom domain. A/B test subject lines. Track opens and replies. Auto-pause when pros