Inspiration
Global supply chains have become increasingly fragile, and construction projects are paying a heavy price. Expeditors, who are responsible for tracking shipments, monitoring geopolitical events, and flagging risks, can no longer keep up manually with the overwhelming volume of changes happening across dozens of suppliers and locations. We were inspired to create a solution that would give expeditors the ability to see problems before they start—transforming reactive fire-fighting into proactive risk management.
What it does
ChainGuard AI is an intelligent supply chain sentinel powered by specialized AI agents. It monitors scheduling and global supply chain risks across construction projects, tracks disruptions by location and severity in real time, and generates clear, actionable insights from complex data. Key features include:
Automated risk summary reports generated with AI Global location-based risk intelligence with geopolitical risk scores Interactive global risk visualization map showing where delays or disruptions are likely Transparent agent reasoning with full logging of every AI action
How we built it
Built on IBM Cloud, ChainGuard AI leverages a multi-agent architecture deployed through IBM watsonx Orchestrate:
Four autonomous agents: Scheduler agent (analyzes delivery timelines), Political risk agent (evaluates global events using integrated web search), Reporting agent (creates comprehensive risk reports), and Assistant agent (handles user interactions) Knowledge sources: IBM Db2 database for structured procurement data and integrated search for real-time global insights Frontend: Intuitive web application for users, Streamlit-based interface for developer testing Backend: FastAPI implementation with chat endpoints, deployed on IBM Cloud Code Engine for scalable, serverless execution Coordination: IBM watsonx Orchestrate enables agents to reason together through multi-agent workflows
Challenges we ran into
Managing the complexity of coordinating multiple specialized agents while ensuring they work together seamlessly was a significant challenge. We also needed to handle fast-changing data from diverse sources—both internal procurement databases and external real-time global events—and present it in a way that was immediately actionable for expeditors under pressure. Accomplishments that we're proud of We're proud of creating a system that makes AI decision-making transparent and trustworthy. The Thinking Logs feature serves as a built-in explainability layer where every AI decision is traceable, showing how data was analyzed and what reasoning led to each output. We've also successfully transformed overwhelming complexity into simple, visual insights through our interactive risk map and automated reporting.
What we learned
We learned the power of agentic AI when specialized agents work together with access to both internal and external knowledge sources. IBM watsonx Orchestrate's transparent multi-agent coordination proved invaluable, and the modular approach with skills and connectors gave us the flexibility to adapt quickly to changing requirements. We also discovered that explainability isn't just a feature—it's essential for building user trust in AI-driven decisions.
What's next for ChainGuard AI
We plan to expand ChainGuard AI's capabilities to cover more types of supply chain risks, including weather events, economic indicators, and supplier financial health. We also want to add predictive analytics that can forecast disruptions weeks or months in advance, and integrate direct communication tools so expeditors can automatically notify stakeholders when risks are detected. Our goal is to make ChainGuard AI the comprehensive command center for global supply chain risk management in construction and beyond.
Log in or sign up for Devpost to join the conversation.