Keel | Devpost

Problem Statement

Capital projects in energy, utilities, and heavy industry depend on long-lead equipment whose delays can halt construction and cause costly schedule overruns. However, delivery risk assessment is largely manual, fragmented, and reliant on expediters monitoring supplier updates and external events. As projects grow across multiple suppliers and countries, this approach becomes unscalable, causing critical risks to be identified too late and resulting in idle crews, financial losses, and delayed project completion.

Proposed Solution

Keel is an AI-powered autonomous risk analyst that continuously monitors long-lead equipment deliveries by combining project schedules, supplier data, logistics information, trade updates, and global news. It intelligently determines whether external events threaten specific shipments, prioritizes high-risk items, and generates evidence-backed risk assessments. Routine findings are logged automatically, while material, costly, or uncertain risks are escalated to human reviewers, enabling proactive decisions before project schedules are impacted.

How We'll Build It

Step 1 – Build the Brain Develop Keel's AI reasoning engine that prioritizes critical equipment, selects the appropriate risk lenses, and determines whether external events pose a real threat to specific shipments.

Step 2 – Connect the Data Integrate project schedules, supplier information, news, logistics updates, and trade regulations. We'll begin with simulated data to validate the system before connecting to live data sources.

Step 3 – Build the Human Review Portal Create a dashboard where project teams can review AI-generated risk assessments, examine supporting evidence, and approve, reject, or request further investigation before any costly action is taken.