PropPilot

Inspiration

Real estate brokers are overwhelmed with inbound calls from potential investors, often missing opportunities while juggling property searches, client meetings, and market research. We envisioned an AI-powered platform that could handle initial client interactions 24/7, automatically capture lead information, schedule meetings, and intelligently search for investment properties - all while the broker focuses on closing deals.

What it does

Our Real Estate AI Agent Platform is a comprehensive solution with three core capabilities:

AI Phone Agent - Powered by Retell AI, our voice agent handles inbound calls from potential investors, asks qualifying questions, captures contact information, and automatically schedules meetings on the broker's calendar.
Intelligent Property Search - Using NVIDIA Nemotron AI, our system searches across Zillow, Google Maps, and other data sources to find investment properties matching client criteria, complete with AI-generated investment analysis.
Unified Dashboard - A modern web interface where brokers can review call transcripts, track lead conversion, monitor property searches, and manage their entire investment pipeline.

How we built it

We architected the platform as a monorepo using Turborepo, enabling code sharing while keeping services specialized:

Frontend: Vite + React 18 + TypeScript + Tailwind CSS Calling Agent Service: Node.js + Express + Prisma ORM, Retell AI API Searching Agent Service: Python + FastAPI + Prisma, NVIDIA Nemotron AI

Challenges we ran into

Making a clean communication between the Vite frontend and two separate backend services, using Retell AI and Nemotron AI, displaying live call transcripts as the conversation happened, converting free-form conversation into structured customer records, scraping data from sources.

Accomplishments that we're proud of

Creating a clean frontend to display, learning new frameworks, effectively using Retell AI

What we learned

Monorepo Architecture, AI Integration, Webhook Handling, Team Coordination.

What's next for Property

We want to expand and make the AI Agent more efficient for more complex situations. We also want to add more complexity by integrating a Google Calendar booking system, use NVIDIA Nemotron better, and overall make the entire application more useful for real estate agents.