Descriptions:
Nate Herk builds a complete outbound lead qualification system using n8n and Vapi that automatically calls new form submissions, qualifies them with a conversational AI voice agent, and logs structured call data to Google Sheets—all without human involvement. The video opens with a live demo: a Vapi agent named Elliot calls a mock prospect named Richard, works through a qualification script covering intent, budget (the lead mentions $5–10K), timeline, and prior automation experience, then wraps the call cleanly.
The tutorial covers every layer of the n8n workflow: a form submission trigger, an HTTP request node that calls Vapi’s API with dynamically injected lead variables (name, phone, company, request), and a polling step that waits for call completion before writing structured output to a spreadsheet. Herk spends significant time on the Vapi system prompt, explaining how to structure conversation flow, define required qualification topics, add error handling for wrong numbers and bad timing, and configure the predefined end-call function—noting that Vapi’s older summary and success evaluation fields are now deprecated in favor of structured outputs.
Key takeaways include prompt design principles for voice agents, how to think about iterative refinement in production by monitoring call logs, and how to architect workflows that capture meaningful data downstream. Free templates and resources are included, making the system straightforward to replicate.
📺 Source: Nate Herk · Published January 12, 2026
🏷️ Format: Hands On Build
