spruvly
Pablo Serrano

Cold-email personalization at scale

Apollo → n8n → GPT-4 generates per-lead personalization based on LinkedIn + company news → Smartlead.

Open rate 41%, reply rate 8.5%, on 800 emails/week.
0

Problem

Outbound team blasted generic emails — reply rate was 0.8% and the domain reputation kept eroding. Cold-email needed to actually feel cold-emailish on the prospect's side.

Flow

Apollo CSV → n8n splits → enrichment in parallel → GPT-4 per-lead prompt → Smartlead campaign API → reply webhook routes hot replies to AE.

Cadence

Mon-Thu, 200/day, ~800/week

What I built

  • Apollo pull of ICP leads with role + company tags
  • Parallel enrichment: LinkedIn snippets + last 3 company news items
  • GPT-4 generates first-line + relevant value-prop per lead
  • Smartlead drip with reply-detection and out-of-sequence routing

Limits

  • No regulated industries (healthcare/finance) — gated by domain blocklist
  • Send-volume capped at 800/week to protect domain reputation
Cold-email personalization at scale

AI tools used

  • OpenAI API
  • n8n
  • Zapier
Pablo Serrano
@pablo.serrano
View profile