A building had 30 residents getting their bills through WhatsApp voice notes and paper under the door. I automated the whole thing. Now nobody calls, nobody argues, nobody chases anyone.
Took me less time to build than one of their meetings lasted.
I deployed a low-latency Voice AI agent that handles front-desk phone calls, end-to-end real calls, real calendar writes, not a static chat demo.
The architecture generalizes to any business dependent on live phone calls and scheduling (clinics, salons, legal practices) anywhere a missed call means a lost customer.
The Production Stack:
• Voice Layer: Vapi + Groq (Llama-3.1) for sub-200ms token generation.
• Logic Gateway: Async FastAPI backend managing structural data validation.
• Workflow Sync: Decoupled n8n subflows to protect core server threads from third-party API downtime.
• Logs & Observability: PostgreSQL persistent decision log with a non-developer dashboard.
If your front desk runs on phone calls and you want to systematically protect your revenue lines from dropped after-hours customer calls, let's talk.
Drop a comment or send me a direct DM to discuss an automation blueprint for your business.
A calendar event → a WhatsApp reminder → a log in Google Sheets.
I built this personal demo in n8n using Google Calendar, Notion, WhatsApp Cloud API and Google Sheets.
The workflow checks upcoming appointments, looks up the recipient, checks the reminder history, and sends the message when it's due.
The video shows a test event, the workflow running, and the message arriving on my own WhatsApp account. This is a demo, not a client deployment.
I'm available for small n8n integration projects, particularly connecting a lead source to an existing CRM with validation, duplicate checks and error logging.
If your agency has a small integration task to hand off, feel free to message me.