AI Solutions · 2026
THADAM
Brief
A lead-generation system that finds businesses in a niche, works out each one's pain point from what's public about them online, and drafts a personalised email about the fix. It had to reach the businesses that actually need help without ever reading like spam.
Materials
- Python
- SQLite
- n8n
- SMTP / Resend
Constraint
Cold outreach reads as spam on sight, and a business that smells a template deletes it by the second line. Nothing in the message could be invented: no fabricated metrics and no guessed familiarity. And nothing could send itself, every message needed a named human to approve it before it left the building.
Form
Each business is assessed against evidence from what's public about it online, like its website and reviews, and only a real finding earns a draft. The draft opens with that one sourced fact and asks for a 20-minute audit; a validation pass rejects any generic pitch, invented percentage, or fake familiarity before a person sees it. A local, read-only console lets a single operator scan every business's stage and severity and approve a draft in one click, and that click is the only action in the whole system that can cause a message to leave.
Outcome
Live and in use. Every email is read and approved by a person before it sends, and most businesses it assesses get nothing, because there is no problem to point to. One client replied "thanks for flagging this", joined a call, and went ahead with a tailored plan.
Next