Building AI Agents - The AI Agent Starts Building the Business Episode #7

In Episode 7, Hermes starts doing real work inside the BuildWithPXL AI infrastructure stack. We move beyond planning and begin wiring together the first user-facing AI experience for the platform. Hermes helps build the initial waitlist flow, connect the app to a private n8n workflow, create a Lead Agent demo users can try, and tighten the Cloudflare permissions needed to protect the whole setup. The goal is simple: show what it looks like when an AI agent is not just answering questions, but actually helping operate and build production infrastructure. In this episode, we work through: setting up the initial waiting list experience creating the Lead Agent demo flow connecting the frontend to a private n8n workflow adding controlled lead-search fields for safer user input expanding Cloudflare API token scope for tunnel and access work protecting n8n behind Cloudflare Access instead of exposing it publicly wiring callback logic so n8n can return enriched lead results to the app polishing the demo result modal so users can see the value clearly Chapters - 00:00 - Intro 01:24 - Initial Prompt Prep 07:04 - Initial Webapp Build 11:15 - Fixing Supabase Migration Issue 14:31 - Revisions on the landing page 19:35 - Further Refinement 21:07 - Photoshop elements 24:19 - Handing off elements to Atlas 26:12 - Image elements added to landing page 28:37 - N8N workflow setup 33:08- Debugging N8N issue 37:12 - Solved N8N Issue 39:42 - Prompting for Leads Agent tool Setup 44:26 - Checking out mock setup of Leads Agent Tool 47:22 - Updating scope of Cloudflare API key 51:37 - Cloudflare scope added and verified 54:51 - Updating APIFY api key 57:04 - First test with Leads Gen Tool 59:59 - Success on Leads Gen Tool Implementation 61:01 - Revising modal layout 62:50 - Outro This is where the stack starts to feel alive. The app, Cloudflare, n8n, Supabase, and Hermes all begin working together as one connected system. We are getting close to the final production AI infrastructure setup. After this, the main focus is bringing the production VPS online, matching the staging architecture, and locking everything down for the full BuildWithPXL tutorial series.