Real Estate AI Property Assistant

Business Problem

Real estate agents often search property spreadsheets manually to answer questions about price, location, bedrooms, property type, availability, and features.

Solution

Built an automated ingestion pipeline that monitors property files, creates embeddings, stores listings in a vector database, and allows users to search property inventory conversationally through an AI agent.

Workflow

Key Features

• Automated property spreadsheet ingestion
• Embedding and vector storage for listings
• Conversational property search with RAG
• Grounded recommendations from inventory data

Tech Stack

Architecture / System Flow

Google Drive updates trigger ingestion and embedding into Supabase, then user queries are answered through RAG retrieval and an AI property assistant.