← Easy USA projects

An AI sales research and outreach system

I built a system that connects buyer research with outreach, follow-ups, and a dashboard for tracking the work.

From research to outreach

A scheduled Claude Code routine researches potential buyers each weekday, reads their websites, and writes personalized email openings. The routine submits prospects to a protected API, where backend JavaScript validates the records, builds emails, and handles sending.

Keeping the work in one place

The backend and dashboard run on Vercel, with Neon Postgres storing prospects, buyer categories, drafts, sent messages, follow-ups, replies, and activity logs. The dashboard brings those records together so the outreach can be reviewed and followed over time.

Connecting to existing tools

Gmail’s API sends from the company sales account and checks replies and bounces. Google Sheets syncs the tracker and research instructions. A read-only Shopify integration checks existing customers and supports order attribution.

Scheduling the follow-through

GitHub Actions triggers follow-ups, tracker syncs, reply checks, and Slack recaps. Vercel provides a daily backup trigger. Code pushed to the main branch deploys automatically.

Where AI fits

Claude handles research and personalized openings; the application handles validation, records, integrations, and sending. Building those parts together turned a research routine into a repeatable workflow with a history of what happened.

Part of the larger project

Modernizing Easy USA →