Inspiration
PropertyAI is an India-first, AI-assisted real estate decision engine designed to help buyers, families, and investors make rational, risk-aware property decisions in a market dominated by misinformation, emotional buying, and opaque pricing.
Unlike listing platforms that optimize for leads, PropertyAI optimizes for buyer safety, capital protection, and long-term livability.
The system combines deterministic data signals (pricing, hospitals, schools, flood risk, air quality) with LLM-based reasoning, while enforcing strict numeric guardrails so the AI never produces unrealistic or fear-driven conclusions.
What it does
Core Problem We Solve The Indian Real Estate Problem No transparent pricing Registry data is fragmented Comparable transactions are rarely available Circle rates are unreliable Critical infrastructure is ignored Hospital access is not evaluated Flood risk is discovered only after purchase Commute assumptions are misleading Advice is biased Brokers push deals Portals push inventory “Reviews” are promotional Tier-2 / Tier-3 India is misunderstood Western risk models incorrectly mark most India as “unlivable” Buyers are either falsely reassured or excessively scared
What PropertyAI Does Differently PropertyAI answers one core question: “Given my price, location, and intent — is this property actually a good decision for me?” It does this using a hybrid decision architecture: Deterministic signals (facts) AI reasoning (context) Guardrails (discipline)
How we built it
USED GEMINI MODEL as agent which gets all data and reasons on it, google maps api for data collection, antigravity for ui designing User Input ↓ Location Resolution (Geocode / Cache) ↓ Region Classification (Tier 1 / Tier 2-3) ↓ Signal Generation (Parallel) ├─ Pricing Intelligence ├─ Hospital Accessibility ├─ Air Quality (WAQI) ├─ Flood Risk ├─ School Ecosystem ├─ Commute Stress (approx) ↓ Numeric Risk Scoring (India-aware) ↓ LLM Reasoning (Gemini / OpenAI) ↓ Decision Guardrails ↓ Actionable Report
Challenges we ran into
Pricing sqft data still needs to be extracted and collected from various sources for more accuracy
Accomplishments that we're proud of
What Makes PropertyAI Unique
🇮🇳 India-first logic
🧠 AI with discipline
📊 Honest uncertainty handling
🏘️ Family-centric design
🛡️ Buyer-safety over sales
This is not Zillow. This is not MagicBricks. This is closer to a “personal real estate risk analyst.”
What we learned
The core thing that we have learned is that when a person coming from a different place and goes to a new place and wants to buy a new property he/she always maynot know behind the scenes issues like flooding data or analyzing hospital schools access distances manually with each location is cumbersome. So this tool will help narrow down the choices and look into the needs specifically required instead of buying anything that's available in a budget.
What's next for Property Decision AI
Real-World Use Cases 🏠 Family Buyers Avoid flood-prone plots Ensure school + hospital balance Prevent emotional overspending
📈 Investors Detect liquidity traps Avoid premium mispricing in Tier-2/3 Validate land rates objectively
🧓 Elderly Buyers Filter unsafe medical access zones
🏗️ Developers Identify infra gaps Price realistically Improve project positioning
🏦 Banks / NBFCs (Future) Risk-aware underwriting Safer loan approvals
Log in or sign up for Devpost to join the conversation.