The problem

A real-estate developer targets companies looking for offices and commercial space. The sales team sends generic LinkedIn messages that get 2% replies, and nobody remembers who said what to whom. Mass automation would get accounts restricted — and deserves to.

The solution, step by step

  1. Build a precise target list with LinkedIn Sales Navigator: sector, company size, city, role (general managers, office managers, expansion leads).
  2. For each prospect, collect public signals by hand or via light automation: recent posts, company announcements, growth news — into Google Sheets.
  3. ChatGPT receives the signals and drafts a three-sentence opener referencing one specific, real element — never a fake compliment, never a pitch in message one.
  4. A human reads, edits and sends every message — the AI drafts, the salesperson decides. Ten a day, no more.
  5. An n8n workflow tracks the pipeline in the sheet: sent, replied, meeting booked, and schedules polite follow-ups at D+5 and D+12 for the human to send.
  6. Weekly, feed reply rates by segment back to the model to sharpen next week's angles.
  7. Result and estimated gains

    • Reply rate typically climbs from 2% to 15-20% because every message is about the prospect, not the sender.
    • 2 to 4 qualified meetings per week from thirty minutes of daily effort.
    • A clean conversation history the whole sales team can pick up at any time.