Front Desk: AI Receptionist, Booking & Lead Scoring
Krystian Siek
Front Desk is an AI receptionist and lead-qualification system that runs in production on my own infrastructure.
What it does
Voice receptionist (ElevenLabs), a website chat with a call button, and the same assistant on Telegram.
Answers from the business's own documents with checkable sources, and says "I don't know" instead of guessing.
Real bookings in Google Calendar, with confirmations, reminders and a cancel link.
Lead scoring: the LLM gives sub-scores with quotes from the conversation, and a deterministic policy gate makes the decision. The model suggests, code decides.
One inbox for every channel, live human handoff, Slack / Telegram / email alerts and a signed webhook into any CRM.
Measured on the running system
Median lead decision 3.0 s across 281 real leads.
Policy gate decision in about 7 ms, with no model call.
44/44 held-out support answers right; 13/13 prompt-injection attacks blocked.
Stack: Python, TypeScript, n8n, PostgreSQL/pgvector, Redis, Docker, ElevenLabs.
Like this project
Posted Nov 5, 2026
AI receptionist on voice, website chat and Telegram: answers from your documents with sources, books into Google Calendar, scores leads behind a deterministic policy gate. Median lead decision 3.0 s.
