Descriptions:
Nate Herk argues that drag-and-drop automation platforms like n8n, Make.com, and Zapier are being displaced by agentic AI workflows — and demonstrates exactly what that looks like in practice using Claude Code and Trigger.dev. The core claim is that the bottleneck in automation has never been what systems can do, but how long it takes to build them; agentic tools collapse that gap by letting developers describe outcomes in natural language instead of wiring nodes by hand.
The video’s centerpiece is a live demo: a ClickUp task containing a company name triggers an AI research agent that searches the web, synthesizes competitive intelligence, and posts a structured brief back to the task — all built by telling Claude Code what the desired outcome was, without manually configuring API calls or mapping variables. Trigger.dev’s visual run log is used to show the agent’s tool calls and decision steps in real time, addressing a common concern about observability in agentic systems.
Herk is careful not to dismiss n8n entirely, framing this as an evolution rather than a replacement, and dedicates meaningful time to real limitations of agentic workflows — particularly context drift over long sessions, where agents begin hallucinating or reverting to earlier patterns. His recommended mitigations include shorter focused sessions and maintaining a running project summary file. The video is aimed at automation builders already comfortable with n8n who are evaluating when and how to bring Claude Code-based agentic workflows into their stack.
📺 Source: Nate Herk | AI Automation · Published March 21, 2026
🏷️ Format: Opinion Editorial
