The stack if you want to know

Every automation we build runs on n8n, an open-source workflow automation platform that handles triggers, logic branching, API calls, and integrations in a single visual pipeline.

We chose n8n because it gives us fine-grained control over every node in the workflow without abstracting away the logic. When something goes wrong, we can see exactly where and why, not just that something failed.

Each client's workflows are isolated. Your data doesn't touch another client's workflow under any circumstance.

Our Phone Receptionist is built on VAPI, a voice AI platform designed specifically for building production phone agents. It handles call routing, speech-to-text, LLM inference, and text-to-speech in a single low-latency pipeline.

VAPI's end-of-call webhook fires a structured JSON payload (caller ID, transcript, call duration, and metadata) into our n8n workflow, which then scores the lead, logs it, and fires alerts.

Average latency from speech to response is under 800ms in production. Callers experience a natural conversation, not a robotic pause.

Lead scoring runs through a custom JavaScript scoring engine inside n8n, backed by Groq for fast LLM inference on unstructured inputs like email bodies and call transcripts.

Scoring factors include job title seniority, company size, industry vertical, inbound source, and intent signals extracted from message content. Scores map to three tiers: Hot, Warm, and Cold. These determine routing, follow-up cadence, and alert priority.

For the Automated Receptionist Suite, RAG (retrieval-augmented generation) is used to pull answers from a client-specific knowledge base stored in Google Sheets, ensuring responses are always grounded in your actual business information.