our product logo

SupplyMind AI

Inspiration

During our research, we found that construction companies can track where a shipment is, but they cannot predict whether it will be delayed. Even a small delay in materials like steel, cement, or HVAC equipment can delay the entire construction project. Project managers often spend hours contacting vendors and transporters to find the reason for delays. This inspired us to build SupplyMind AI, an AI-powered platform that predicts delays early and helps managers take action before the project is affected.

What it does

SupplyMind AI monitors every stage of a material, from order approval to final delivery. It predicts the probability of shipment delays using historical supply chain data. Instead of only showing shipment status, it explains why a delay may happen, measures its impact on the project schedule, recommends better vendors or routes, and sends alerts to project managers through the dashboard, email, or WhatsApp.

How we built it

We prepared a construction supply chain dataset and trained an XGBoost machine learning model to predict shipment delays. SHAP is used to explain why the model made a prediction. FastAPI serves the machine learning model, PostgreSQL stores shipment information, and Streamlit provides an interactive dashboard. We also integrated weather and route APIs to improve prediction quality and provide better recommendations.

Challenges we ran into

One of the biggest challenges was preparing a realistic construction supply chain dataset because public datasets were limited. We also had to select useful features, handle missing values, and balance the data for better predictions. Another challenge was designing a dashboard that presents predictions, explanations, and recommendations in a simple and user-friendly way.

Accomplishments that we're proud of

We successfully built an end-to-end AI-powered supply chain prototype that predicts shipment delays instead of only tracking them. We integrated machine learning, explainable AI, project impact analysis, smart recommendations, and an interactive dashboard into one solution. We are also proud of creating a realistic construction supply chain dataset and building a working prototype within the hackathon timeline.

What we learned

Through this project, we learned that building an AI solution is not only about training a machine learning model. Clean data, feature engineering, explainable AI, and a user-friendly dashboard are equally important. We also gained practical experience with XGBoost, SHAP, FastAPI, PostgreSQL, Streamlit, and real-world AI application development.

What's next for SupplyMind AI

Our next goal is to make SupplyMind AI even smarter by adding real-time GPS tracking, IoT sensor integration, demand forecasting, ERP integration, and a mobile application. We also plan to build AI agents that can automatically recommend the best supplier, delivery route, and recovery plan whenever a shipment is at risk of delay.