FrontDesk AI is an end-to-end AI voice receptionist system. A prospect calls or clicks Talk to Aria on the website. Within 3 seconds Aria answers, holds a natural qualifying conversation, checks real-time calendar availability, books a confirmed strategy call, and closes the call. After the call, a post-call modal captures additional context from the prospect which triggers a structured pre-meeting brief delivered to the ANR team.

Five Layers of the System

• Layer 1 — Voice Interface: Vapi handles call routing, real-time transcription, and tool execution. Gemini 2.5 Flash powers Aria’s reasoning and conversation logic.
• Layer 2 — Qualification Engine: Aria evaluates six weighted factors mid-conversation — decision maker authority, budget signals, timeline urgency, email domain quality, use case specificity, and company size — building a lead score in real time.
• Layer 3 — Booking Automation: Mid-conversation, Aria triggers the checkAvailability tool to query Cal.com for real-time slot availability, offers two options to the caller, and fires the bookAppointment tool which routes through n8n to Cal.com to create the confirmed booking.
• Layer 4 — Post-Call Data Capture: After the call ends, a frontend modal pre-fills the caller’s name, email, and problem summary captured during the call. The prospect reviews, edits if needed, selects qualification checkboxes, and confirms. This triggers n8n Workflow 2.
• Layer 5 — Notifications and Persistence: n8n Workflow 2 updates the Supabase lead record with enriched data, sends a prospect confirmation email from ANR Digital Studios, and delivers a structured pre-meeting brief to Aryan and Rohit via email and Slack.

The Full Flow in One Line

Inbound Call → Aria Answers → Qualify → Check Availability → Book on Cal.com → Supabase Log → Slack + Email → Post-Call Modal → Enriched Brief to Team

Technical Stack

Average call to confirmed booking: Under 90 seconds including email delivery.