I built an AI Agent in 2 hours (and got paid $2600)

I built an AI Agent in 2 hours (and got paid $2600)

More

Descriptions:

Automation developer Nate Herk builds a complete student onboarding system live in n8n—the same system a client paid him $2,600 for—and walks through both the business rationale and the technical implementation. The system comprises four distinct workflows: a webhook-triggered welcome sequence that sends an email, posts a Slack notification, and creates a CRM entry in Google Sheets; two daily automated checks for students who haven’t registered; and a human escalation alert sent to the team after five days of inactivity.

Only one of the four workflows uses AI—a language model agent that writes a personalized welcome email once a student completes registration. The remaining three workflows are purely deterministic automations, a deliberate design choice Herk argues makes the system more robust and easier to sell to clients unfamiliar with AI reliability. He uses Postman to simulate webhook payloads during the live build and validates each Google Sheets update on screen.

Beyond the technical build, the video addresses freelancing fundamentals: wireframing client systems before opening n8n, using clear status columns in a Google Sheets CRM to prevent students from slipping through cracks, and why minimizing AI in early client builds reduces scope creep and increases trust. Herk also briefly references his training program for those interested in building a similar automation services business.


📺 Source: Nate Herk · Published December 17, 2025
🏷️ Format: Workflow Case Study

1 Item

People