24/7 Conversational AI Voice Receptionist & Lead Booking Engine (5-in-1 Agency Kit)

Calculator

Step-by-step workflow

1. Provision Low-Latency Webhook Infrastructure on Hostinger Cloud VPS

   Deploy a Dockerized n8n instance on a Hostinger KVM 2 VPS using Coolify or Docker Compose to guarantee sub-200ms webhook response times during live phone calls. Configure SSL certificates and dedicated environment variables so your voice agent never drops mid-conversation due to cold-start timeouts.

2. Configure ElevenLabs Conversational AI Agent & Custom Tools

   Create a new Conversational AI agent inside ElevenLabs, select a low-latency Turbo v2.5 voice model, and define two Server Tools (`check_calendar_slots` and `book_appointment`). Point both tool webhook URLs directly to your self-hosted n8n or Make.com production endpoints.

3. Wire Twilio SIP/Phone Number to ElevenLabs WebSocket Stream

   Connect a local Twilio phone number to the ElevenLabs native Twilio integration or forward inbound TwiML voice streams directly to the ElevenLabs Conversational WebSocket endpoint. Enable caller ID passthrough so the caller's phone number is automatically injected into your webhook payload.

4. Orchestrate Real-Time Calendar & CRM Sync in Make.com or n8n

   Build the webhook receiver in n8n or Make.com to parse the caller's requested date, query Google Calendar or Cal.com for available 30-minute windows, and return natural-language availability back to ElevenLabs within 800ms. Upon booking confirmation, log the transcript and audio recording URL into Google Sheets.

5. Automate SMS Confirmation & HeyGen Personalized Video Follow-Up

   Once the call ends and the post-call webhook fires, trigger an automated Google Apps Script or Make.com router that texts the calendar invite to the lead. For high-ticket medical or real estate leads, automatically invoke the HeyGen API to send a personalized 30-second video welcome from the business owner.

Ready-to-use AI prompts

You are Sarah, the warm, ultra-efficient front-desk receptionist for [Business Name]. Keep every response under 2 short sentences so you sound natural over the phone. Never use markdown, bullet points, or emojis. Your primary objective is to answer basic clinic/office questions, qualify the caller's service need, and book them into an available slot using the `check_calendar_slots` and `book_appointment` tools. Always confirm the caller's full name and phone number before finalizing the booking. If the caller asks for pricing you do not know, state the starting consultation fee and offer to lock in a priority slot.

Analyze the live conversation state and extract structured JSON parameters required for the `book_appointment` webhook tool. Ensure `caller_name` is formatted in Title Case, `caller_phone` is normalized to E.164 format (+1XXXXXXXXXX), `requested_iso_datetime` is converted to the business timezone (America/New_York), and `service_category` matches one of the approved intake types. If any required field is ambiguous, ask a single clarifying question before invoking the webhook.

You are an Executive CRM Analyst. Review the complete ElevenLabs call transcript below and output a strict JSON object containing: 1) `lead_intent_score` (1-100), 2) `appointment_booked` (boolean), 3) `key_objections` (string summary), 4) `sms_followup_draft` (under 160 characters personalized to their exact request), and 5) `heygen_video_script` (a 45-word spoken script welcoming the lead by first name and confirming their upcoming appointment time).