Descriptions:
Stephanie Nyarko demonstrates a fully automated lead generation pipeline built on Claude’s computer use feature, which she calls Claude Cowork. The system targets local businesses — in this walkthrough, wellness spas in Ottawa — and autonomously searches for prospects, scrapes Google review data, identifies business gaps from customer feedback, and drafts personalized cold outreach emails that land in Gmail drafts for human review before sending.
The workflow is built on four components: a Google Sheets database that tracks all leads and contact history, a skill file (a plain-text .md document that instructs Claude on the exact steps to follow), a config file storing personal settings and API keys, and an App Script bridge that lets Claude write safely into Google Sheets. All code can be AI-generated, making the system accessible to non-developers.
Nyarko shows how scaling from 2 to 1,000 leads requires nothing more than changing the target number in the initial prompt. The system eliminates manual copy-pasting from Google Maps, removes the need for a virtual assistant to handle scraping, and maintains a human-in-the-loop review step before any email is sent — a deliberate design choice to preserve quality control. This is a practical blueprint for AI consultants and service businesses looking to systematize outbound prospecting without sacrificing personalization.
📺 Source: Stephanie Nyarko · Published May 27, 2026
🏷️ Format: Tutorial Demo
