Inspiration Manually finding B2B leads for water-reuse systems takes hours of digging through satellite imagery and corporate ESG reports. We wanted to use AI to turn this tedious research into a single click.
What it does An AI-powered prospecting engine that maps approximately 4,000 large Texas commercial buildings in DFW and Austin. It automatically calculates water-capture ROI and uses Gemini API with the Google Maps API key to detect nearby cooling towers and summarize corporate sustainability goals.
How we built it Backend: Python, FastAPI, and SQLite for lightning-fast caching. Frontend: React, Vite, Tailwind CSS, and Recharts. APIs: Google Maps (interactive map & satellite views) and Google Gemini (automated ESG and property analysis).
Challenges we ran into Running complex AI prompts on every click caused UI timeouts. We solved this by engineering a smart caching layer—Gemini only runs in real-time if the building hasn't been analyzed yet, saving the results to our database for instant future retrieval.
Accomplishments that we're proud of Building a tool that successfully synthesizes physical data (square footage), climate data (rainfall), and qualitative data (ESG reports) into one single "Viability Score." with ability to filter as needed.
What we learned We learned how to effectively balance real-time GenAI calls with aggressive database caching to keep the frontend snappy without sacrificing the dynamic power of the LLM.
What's next Improving performance to incorporate all of Texas and then other states.
Built With
- fastapi
- gemini
- google-maps
- html
- python
- sqlite
- tailwind
- typescript
Log in or sign up for Devpost to join the conversation.