How I’m using Claude Code to automate lead sourcing every morning

Claude Code has become one of my most useful workflow tools because it can work across files, run commands, use project memory, and connect to external tools instead of me repeating the same instructions every day. I use CLAUDE.md for persistent rules, Skills for repeatable multi-step playbooks, Hooks for deterministic automation, and MCP when I want Claude Code to work with external data sources and APIs.

When I sit at my desk in the morning, I just open my lead-sourcing workspace, run Claude Code, and ask it to generate a fresh batch of companies that match my ICP, enrich them, clean duplicates, and sort them into outreach-ready lists.

Here is how I have set it up:

• create one project folder just for lead sourcing
• create /input, /output, /research, /scripts, /prompts, /logs folders
• keep raw company lists inside /input/raw
• keep cleaned lists inside /input/cleaned
• keep final outreach-ready CSVs inside /output/final
• keep persona-specific folders like /output/seo-agencies, /output/b2b-saas, /output/healthcare
• create a root CLAUDE.md file
• write ICP rules inside CLAUDE.md
• write exclusion rules inside CLAUDE.md
• write output column rules inside CLAUDE.md
• write file naming rules inside CLAUDE.md
• write “never guess emails” inside CLAUDE.md
• write “mark unknown instead of inventing data” inside CLAUDE.md

This is just a quick overview of how I’ve set this up and how it’s working for me right now. There’s a lot more you can customize based on your niche, ICP, and workflow. If you have any questions or want me to break down a specific part, just drop a comment 👍