AI Real Estate Agent Development by Jacob DzikowskiAI Real Estate Agent Development by Jacob Dzikowski

AI Real Estate Agent Development

Jacob Dzikowski

Jacob Dzikowski

AI Real Estate Agent That Matches Users With The Right Apartments ๐Ÿ 

Overview ๐Ÿ“–

This project is an example of a fullstack development combining an AI agent FastAPI python backend with a custom frontend widget written in javascript. Using Python, FastAPI, and models like OpenAI and Claude, the system guides users through a natural conversation in Polish, collects their housing preferences, and recommends the best-matching apartments from the clientโ€™s database.

Collaboration with Lantriq ๐Ÿค

A Polish real estate company partnered with me to turn their raw listings into a smart digital agent that qualifies leads, recommends properties, and sends contact details directly to their sales team, so that they can connect the lead and close it within seconds, knowing their preferences.
Widget Conversation in Polish And Flat Selection

Key Challenges ๐Ÿค”

Leads leaving the website without sharing their contact information
Unqualified leads reaching out to the real estate company
Capturing high-quality leads without adding friction
Seamlessly connecting the backend to the frontend with the speed of chat to engage visitors

Approach ๐Ÿ”

AI Agent Backend (Python & FastAPI) Developed an AI agent backend in Python and FastAPI that interprets user queries, collects preferences (location, budget, size, parking, avoiding city center, etc.), and scores apartments based on structural data and business rules. The backend connects to the clientโ€™s API to fetch and periodically update available listings.
Matching Logic & Fallbacks Implemented a robust matching engine that ranks offers, handles edge cases when no perfect match is found, and applies clear fallback logic so users always receive relevant suggestions instead of dead ends.
JavaScript Widget Frontend Built a lightweight JavaScript chat widget that can be embedded directly on the clientโ€™s website. The widget provides a smooth conversational interface, sends messages to the backend over secure APIs, and displays personalized apartment recommendations in real time.
Custom Integration & Lead Capture Integrated with the clientโ€™s existing systems using Node.js and REST APIs. The agent saves qualified leads directly into Supabase, ensuring the sales team receives structured contact data and user preferences without manual copying.

Provided Services & Solutions โœ…

๐Ÿ“Œ Fullstack development (Python, FastAPI, JavaScript, Supabase) ๐Ÿ“Œ AI agent development (OpenAI, Claude, Pydantic) ๐Ÿ“Œ Real estate data integration and matching logic ๐Ÿ“Œ Custom website widget and frontend integration ๐Ÿ“Œ Lead capture and Supabase integration ๐Ÿ“Œ Architecture design & technical consulting

Next Steps ๐Ÿ‘ฃ

If you're looking for someone with 6+ years of experience to bring your vision to life, click the link below.
Like this project

Posted May 29, 2026

AI-driven real estate agent for lead qualification and property recommendations, combining fullstack development, Python, FastAPI, and a JavaScript chat widget.