The Problem: Churn happens in the shadows. Disconnected data—failed payments, angry emails, and dropping usage—means Customer Success teams react too late, costing the business thousands in lost MRR.

The Solution: I engineered an autonomous n8n workflow that acts as a 24/7 AI Retention Agent. It ingests real-time signals (Stripe, IMAP, Usage Scans), enriches them with CRM data, and uses GPT-4o + Pinecone Vector DB to predict churn risk.

Instead of just sending alerts, it autonomously executes the perfect retention strategy: Critical Risk: Instantly pages the Exec team via PagerDuty & Slack. 🟡 Medium Risk: Drafts a custom discount email, pausing for Human-in-the-Loop approval before sending. 🟢 Expansion: Auto-generates a PDF report and triggers a Slack upsell alert for Sales.

The Business Impact: 📉 Predicts churn days before a customer cancels. ⏱️ Saves 20+ hours/week of manual CS triage and data hunting. ️ Protects MRR with safe, AI-driven, context-aware interventions.

Tech Stack: n8n • GPT-4o • Pinecone (RAG) • HubSpot • Stripe • PagerDuty