All case studies
Retail · Automation Shipped Nov 2025 · 19-day build

A D2C coffee brand running on zero ops.

Vigor Beans is a 3-person specialty coffee brand. They didn't want to hire ops staff — they wanted the business to run itself between roasts. We automated their entire back-of-house in 19 days.

The Challenge

Three founders, all baristas, none ops people. They were spending 20 hours/week between them on order management, inventory updates, customer email, and shipping coordination. The business was profitable — but only on paper, given their time cost.

What we shipped

  • Shopify ↔ inventory sync (real-time, multi-warehouse)
  • AI customer support inbox (Gorgias + Claude)
  • Auto-generated weekly newsletter (Klaviyo)
  • Roast scheduling agent (demand-forecast driven)
  • Subscription churn-prevention bot

Stack

  • n8n · custom Node webhooks
  • Shopify · Gorgias · Klaviyo · ShipStation
  • Claude Sonnet 4 for support + email
  • Supabase · Resend
  • Cloudflare Workers
$0Ops headcount needed
19hReclaimed / week
−42%Subscription churn
19dFrom kickoff to live

The unsexy 80%: integration, not models

People expect AI case studies to be about clever models. The Vigor project was 80% integration plumbing and 20% AI. The win came from making 7 separate tools — Shopify, Gorgias, Klaviyo, ShipStation, QuickBooks, Slack, Google Sheets — talk to each other reliably.

"We thought we'd need to hire an ops manager. Instead Growvate built one. It's been running for four months and we forgot it existed — in the best way." — Jordan K., Co-founder, Vigor Beans

The five automations that did the work

  1. Inventory sync. Every Shopify sale decrements inventory across both warehouses, triggers reorder thresholds, and posts daily Slack summaries.
  2. Support inbox. 86% of customer emails (mostly "where is my order") are auto-answered with order-aware context pulled from Shopify and ShipStation.
  3. Newsletter generation. Every Sunday, a Claude prompt drafts that week's email — featured roast, customer story, brew tip — then posts to Klaviyo as a scheduled send. A founder approves in 60 seconds on Sunday night.
  4. Roast scheduling. The system pulls 30-day demand trends and pre-orders, then generates a roast schedule the founders approve each Monday.
  5. Subscription save bot. When a subscriber clicks "cancel", they get a Claude-powered short conversation offering a 1-month pause, a different roast, or a discount. Churn dropped from 11% to 6.4%.

What we'd do differently

We'd build the founder approval surface first. Vigor's founders are hands-on — they wanted to see what the automations were doing, not just trust them. The first version we shipped without an approval dashboard, and they didn't trust it. Two days of building a one-screen "what AI did today" dashboard fixed that completely.

"Our business now feels like a software product. We focus on roasting and customers. The boring stuff runs itself."

— Jordan K.Co-founder, Vigor Beans
Next case study →

Madrah.com — an AI tutor that kids actually finish

Read next case