An email engine that pauses itself before it hurts the domain
A sending platform on AWS SES that verifies every address, warms up slowly and stops on its own when complaints rise.
- System and automation
- Live since September 2026
At a glance
Client Kefey
Client
Kefey
Industry Luxury beverage brand, D2C and wholesale
Industry
Luxury beverage brand, D2C and wholesale
Time to build 8 weeks to production
Time to build
8 weeks to production
Year 2026
Year
2026
Status Live since September 2026
Status
Live since September 2026
Every case on this site answers the same four questions, so you can compare them.
Before
Kefey paid per subscriber for SaaS email tools. Linking the CRM to the main domain sent every company email into one dashboard, and replies went out from a generic address.
What it does now
Lists come in by CSV. Every address is verified and sorted into deliverable, risky or undeliverable.
Contacts are segmented, and each campaign is built in a visual sequence editor.
A paced queue sends through AWS SES within daily warmup caps, with one-click unsubscribe.
Bounce and complaint events come back from SES and update the suppression list.
Any complaint over the threshold pauses sending. Bounces pause it after 50 sends and 3 bounces.
Replies are routed back by signed reply address, auto-replies are filtered out, and engaged contacts sync to GoHighLevel.
Where a person decides
Before launch, an operator decides whether risky addresses are excluded or sent anyway. Kefey signs off every template and the wholesale sequence. Stop and send-now are manual buttons.
Result
97.7%
Delivered during warmup
1,392 of 1,425 sends, 31 Aug to 11 Sep 2026
Spam complaints
Across the whole warmup
165 / 165
Tests passing
At the implementation milestone, 18 suites
How we built it
Designed the database first: contacts, segments, templates, campaigns, sends, events and suppressions.
Connected an address verification service and re-check any result older than 30 days.
Wrote the warmup schedule: 15 sends a day at the start, 320 a day by the end of the second week.
Wired SES and SNS events to the suppression list, plus the auto-pause rules.
Built reply routing, auto-reply filtering and the GoHighLevel sync.
Built analytics, deliverability and campaign dashboards, and a test suite before going live.
Stack
- Next.js
- React
- Supabase Postgres
- AWS SES
- AWS SNS
- GoHighLevel
- Vitest
- Vercel
From the project
More work like this
One inbox for every inquiry a Dubai studio gets
A distributor's back office rebuilt to sell nationally
229,314 contact records audited down to the 193,771 that count
Your case study starts with an audit.
One to two weeks inside your operations. You get a map of where time and money leak, the return on each fix, and a fixed scope before we build anything.
Most businesses we meet have paid for an AI tool nobody uses. We start from your numbers, so the first build is chosen by payback.
Zain Khan and Anwar Sheikh
Founders, OutGrow
Outgrow
Intelligence
Studios.