Small businesses face countless hidden risks, from signing unfavorable leases to partnering with the wrong people. These decisions often have long-term consequences, yet most small business owners lack the tools to evaluate those risks, especially when considering shared or co-ventured spaces.

We wanted to build something that helps real people, not just investors, make smarter, safer decisions in moments that can define the future of their business.

Co-SQRD is Co-venturing and risk analysis made easy, an AI-powered tool that helps small business owners:

• Match with the right co-venturer based on trust, compatibility, and fit
• Evaluate location risk scores using property characteristics such as crime rate, foot traffic, and appreciation rate
• Understand operational fit and potential conflicts before signing a lease

• Frontend: Developed a clean, user-centered interface using Streamlit
• Backend: Integrated a Python-based TensorFlow model to calculate risk scores (Risk Score is calculated based on crime rate, appreciation rate, foot traffic, proximity to infrastructure and neighborhood reputation. These features were chosen through market research and a feature correlation matrix).

• Real estate APIs were either paywalled or too broad, and no niche-free options were available for our use case
• We had to collect and clean real commercial addresses for our dataset manually due to the lack of publicly available datasets for commercial property risk analysis
• With limited time, we couldn't build or integrate a full machine-learning model for risk-scoring
• We faced limitations in customizing the interface due to Streamlit's restricted support for CSS styling
• MongoDB's server being down resulted in difficulties seamlessly storing user and property data

• Identified and tackled a real-world gap in risk assessment for commercial properties that directly impacts real people
• Developed a functional prototype with realistic, insightful data and a user-friendly interface
• Defined a clear product vision with strong future potential for scalability in the real estate space

• You don’t have to build everything to build something meaningful
• Focused execution, even with time limits, can make a niche idea come alive
• Even simple features like foot traffic and nearby businesses can reveal powerful insights when framed well
• Limited data paired with strong design choices can power meaningful and functional demos

• Add a call and messaging system powered by an AI agent
• Expand to more cities and add real-time business profile-matching
• Incorporate operational fit into the property and risk evaluation model
• Integrate with commercial real estate APIs for live listings and smarter scoring