Eddy E

Sales & Ops Automation Systems — n8n + Claude

Aug 2026 - Present • 2 mos

The business already had automations. That was part of the problem. Sales, marketing, operations and client management had accumulated disconnected Zaps and quick fixes, with nobody having a clear picture of how the system actually worked. I rebuilt the automation layer as an automation estate, not another collection of isolated workflows. Systems architecture: → n8n workflows for complex automation logic → Make.com scenarios where visual orchestration fit better → Claude-powered classification and decision steps → AI-assisted drafting workflows → GoHighLevel lead capture and CRM automation → Lead routing and pipeline management → Conditional branches and filters → Error handling and recovery paths → Cross-platform reporting and notifications → Data synchronization between systems → Existing automation audit and repair → Documentation and system map The result is the kind of infrastructure that becomes easier to manage as it grows.

N8n + Claude Workflow: Webhook to CRM Automation

May 2026 - Present • 5 mos

This was less about building an AI system from scratch and more about removing the pile of repetitive work sitting between a lead form and a sales team. The business was receiving leads through forms and webhooks, but every submission still required someone to read it, decide what it meant, draft a response, enter the information into the right place and keep the process moving. I rebuilt that middle layer in n8n with Claude AI automation. Workflow components: → Webhook-based lead capture → Claude-powered lead classification → AI-generated reply drafting → CRM and Google Sheets routing → Automated email processing → Human approval before AI-written responses are sent → Error handling and workflow documentation Each submission is captured, passed to Claude for classification and reply drafting, then routed according to the required destination—CRM, Google Sheets or email. I deliberately kept a human approval step before any AI-written response leaves the system. The result is a useful distinction between automation and autonomy: the system does the repetitive thinking and data movement, while the team still controls what actually gets sent.