AI-powered voice assistant for real estate that automates inbound and outbound calls, qualifies leads, and provides full control through an intuitive admin and user interface.
Duration
6 months
Platform
Web Application
Project Type
AI Integration
Industry
Real Estate
Qualifies inbound real-estate leads 24/7 with natural voice conversations, cutting response time from hours to seconds.
QalifAI set out to solve a real constraint in real estate, one where existing tools were either too rigid, too expensive, or simply not built for the way the business actually operated.
Our team partnered closely with stakeholders to map the workflow end-to-end, identify the highest-leverage improvements, and design an experience that felt effortless to the people using it every day.
The result is a product engineered for scale from the first sprint and not a prototype that needs rebuilding to grow.
We translated the challenge into a focused set of capabilities, each one measured against a concrete business outcome rather than a feature checklist.

The core Nest.js backend handles the AI voice pipeline with Puppeteer for call detection and OpenAI + ElevenLabs for natural conversations.
Real-time WebSocket-powered call logs and live updates for admins to monitor ongoing and past calls.
Centralized dashboard showing usage, AI metrics, and system health with full control over brokerages and users.

AI Voice Engine & Call Flow
Runs 24/7 using Puppeteer to detect and accept new calls
AI Voice Engine & Call Flow
Streaming STT/LLM/TTS pipeline with ElevenLabs and OpenAI
Real-Time Call Monitoring
Live transcription of conversations
Real-Time Call Monitoring
Call status updates with timestamps
Admin Stats & Control Panel
Total users, brokerages, call count and success KPIs
Admin Stats & Control Panel
Time-filtered metrics (Today, Week, Month)
AI Script & Rule Customization
Conditional rule-based logic for call routing
AI Script & Rule Customization
Step-by-step conversation script builder
User Demo & AI Testing Panel
Voice status indicators (Ready, Muted, Connected)
User Demo & AI Testing Panel
Live transcript of test conversations
Call Routing Rules
Custom call flow trees with fallback paths
Call Routing Rules
Route to human agent or CRM after AI qualification
Area-Based Coverage Rules
Zip-code and state-level filtering logic
Area-Based Coverage Rules
Supports service area-based restrictions
CRM Sync & Post-Call Handoff
CRM integration with FUB and others via webhook/API
CRM Sync & Post-Call Handoff
Lead enrichment and note generation using AI
Every screen was shaped around real tasks, reducing friction, surfacing what matters, and staying calm under pressure.

The core Nest.js backend handles the AI voice pipeline with Puppeteer for call detection and OpenAI + ElevenLabs for natural conversations.

Real-time WebSocket-powered call logs and live updates for admins to monitor ongoing and past calls.

Centralized dashboard showing usage, AI metrics, and system health with full control over brokerages and users.
A clean, layered architecture keeps the system maintainable and fast, separating interface, logic, and data so each can evolve independently.
Frontend
React · Tailwind CSS
Backend
Nest.js
Database
PostgreSQL
AI & Voice
Tailwind CSS · Puppeteer · FFmpeg · OpenAI API · ElevenLabs API
Realtime
Socket.IO · WebSockets
Integrations
JWT Auth


Qualifies inbound real-estate leads 24/7 with natural voice conversations, cutting response time from hours to seconds.
Runs 24/7 using Puppeteer to detect and accept new calls
Handoff logic for transferring to humans or external CRMs
Error tracking and retry logic
A deliberate selection of tools, chosen for reliability, speed, and the people who maintain them.
Let's talk about how Devnixa can take your product from idea to a polished, production-grade system.