Solution & Approach: I built a 52-step AI automation system running on self-hosted n8n that monitors 4+ tech publications hourly through RSS feeds, extracts clean content via Firecrawl even from JavaScript-heavy sites, and processes 400-500 articles monthly. The system applies double-filter quality control: first checking content relevance to their product space, then verifying publication authority to focus only on outlets with real audiences. It extracts author names, validates business emails, then enriches contact data through
Apollo to pull work emails, LinkedIn profiles, and professional backgrounds. Every contact logs to a Google Sheets database with three anti-spam checks: 90-day contact cooldown, no-contact flags, and Gmail draft detection. The AI generates hyper-personalized outreach using GPT-4o-mini, referencing the specific article they published, their professional background, and a contextually relevant collaboration offer. Automated follow-ups trigger after a week with no response, and everything tracks to prevent double-contact.