← All case studies
AI AUTOMATIONCOMING SOON
AI-Personalized Cold Email Outreach Engine
Hero Metric
Self-funded our agency pipeline with a single n8n workflow
The Brief
Neotix Labs needed inbound to stop being our only growth channel. Manual cold email was a non-starter at agency volume. Personalization breaks at scale, and copy-paste outreach gets you blocked. We needed an engine.
The Build
Stack: n8n, Apollo, Claude Haiku, and Gmail (warmed inboxes). Apollo scrapes the lead list. n8n fans out into per-lead branches. Claude reads the prospect's site or LinkedIn and writes a one-line opener referencing something specific. Their last launch, a podcast appearance, a hiring post. Send is throttled to 30 per day per inbox across 3 rotating domains. Replies route to a single shared inbox with priority scoring.
The Result
First 30 days: 1,200 sends, 18% open rate, 4.2% reply rate, 14 booked discovery calls. Two became signed retainers within 60 days. The build paid for itself before we shipped the first invoice.
What We'd Do Differently
Started with Claude Sonnet for the personalization step. Massive overkill. Haiku produces near-identical openers at one eighth the cost. Next iteration: replace Apollo with a self-built scraper to drop the per-lead cost floor.