Company

Floorplan.ai transforms user inputs and rough sketches into intelligent, editable floor plans using AI-powered spatial planning. The prototype makes professional-grade layout design accessible to anyone who needs quick spatial visualizations without CAD expertise.

The Challenge

Studies show visualization improves homeowner comprehension by up to 80% and significantly enhances decision accuracy1. Yet traditional CAD software requires specialized knowledge and extensive time investment, making these powerful tools inaccessible to most users. The technical challenge was clear: build an AI system that could convert natural language descriptions like 'open concept kitchen' or 'master bedroom with ensuite' into mathematically accurate, editable floor plans in real-time. The prototype needed to prove that AI could understand human design intentions and deliver professional-quality layouts without requiring any specialized expertise.

The Solution

Ricardo Arcifa, a full-stack developer and A.Team builder, tackled this challenge at A.Team's GenAI Hackathon by creating an AI-powered prototype that could interpret spatial design intentions. He built a React-based interface that translated natural language descriptions into structured layouts using OpenAI's GPT API, with a Node.js backend processing AI responses into renderable floor plan components. The key innovation was treating GPT as a semantic interpreter that could map human descriptions to actual spatial relationships and constraints. Working in a developer-led team, Arcifa architected a modular system using TypeScript that maintained real-time responsiveness while users modified their designs. The entire solution was deployed on Vercel's serverless architecture, proving that AI-powered design tools could be both powerful and lightweight.

Arcadium 1