Properti : AI Chatbot for Personalized, Always-On Property Discovery

Real Estate

Properti (Australia) – AI Chatbot for Smart Property Discovery

Properti revolutionized the real estate journey by integrating our AI-powered chatbot to guide users through property searches and inquiries. This conversational assistant helps potential buyers and renters explore listings based on budget, location, lifestyle needs, and current market availability. It connects seamlessly with MLS databases and CRM systems, delivering instant responses and follow-ups across multiple touchpoints. The AI learns user behavior over time, personalizing listings and increasing qualified leads for agents. With 24/7 support and natural language interaction, Properti’s virtual agent is turning casual browsers into serious buyers.

Project Overview

Client: Properti (Australia-based proptech company serving realtors and real estate marketplaces)
Challenge: Users struggled to find relevant listings quickly and agents faced lead fatigue
Goal: Build a smart AI chatbot to:
- Guide users to the right listings using conversational inputs
- Personalize recommendations through behavior-based learning
- Sync with property databases and CRMs for lead qualification and follow-up
Team: 7 (2 NLP Engineers, 2 Frontend Devs, 2 Backend Devs, 1 Product Manager)
Timeline: 4 months (Prototype → User Testing → Live Deployment)

“GenX helped us turn passive browsers into active prospects. The AI assistant is now our frontline agent—smarter, faster, and always available.”

CEO & Co-Founder, Properti

The Challenge

Critical Pain Points:

Property seekers dropped off due to irrelevant search results and complex filters
Agents were overwhelmed by low-quality inquiries with little buyer intent
There was no system for proactive, personalized engagement beyond static search pages

Technical Hurdles:

Handling vague or multi-intent user queries (e.g., “pet-friendly near beach with 3 beds”)
Building behavior-driven personalization into conversation flow
Syncing listings with dynamic market updates and ensuring real-time availability

Tech Stack

Component	Technologies
NLP & Chatbot Engine	Rasa, BERT, GPT-3.5 Turbo, Dialogflow CX
Backend & Lead Sync	Node.js, Firebase, PostgreSQL
MLS & CRM Integration	REST APIs, GraphQL, Webhooks
Personalization & Scoring	Scikit-learn, XGBoost, Behavior Analytics Engine
Frontend & Widget Interface	React, Vue, WebSocket
Monitoring & Notifications	Datadog, Mixpanel, Twilio SMS & WhatsApp API

Key Innovations

The chatbot used user preferences and lifestyle data to recommend relevant properties. It synced with MLS and CRM systems to enable instant responses and follow-ups. 24/7 interaction turned casual interest into serious leads.

Intent-Aware Conversational Search

Transformed vague inputs into real-time filtered listings

Result: 39% increase in listing click-through rate within 7 days

Lead Intent Prediction

Assigned buyer readiness scores based on activity and query behavior

Result: 52% improvement in agent conversion productivity

Proactive Follow-Up Automation

Nudged users back into the platform with personalized updates

Result: 27% more returning users and confirmed viewings

Our AI/ML Architecture

Core Models

Conversational Search Engine:
- BERT-based intent recognition with fuzzy matching and query expansion
- Translates user language into structured filters (bedrooms, suburb, price range, etc.)
Personalization & Lead Scoring Layer:
- Learns user behavior over time to recommend properties and score lead intent
- Triggers contextual nudges and proactive listing suggestions
CRM Integration & Follow-Up Engine:
- Syncs agent calendars and contact pipelines
- Notifies agents once premium leads exceed activity benchmarks