Inspiration
Humanity spends a lot of time constructing, and a lot of time is just spent planning, gathering crew, and finally creating long lasting structures. For each individual worker, the plan means nothing, but the worker means a lot to the plan.
What it does
We designed SiteMind as an AI-powered construction state intelligence system. it uses information about procurement, delivery, structure inspections, meetings, and uses them together to act as a system planner. When a new event occurs, SiteMind: Extracts and standardizes the event. Updates the project-state graph. Identifies affected tasks and downstream dependencies. Evaluates project risk and potential delays. Generates role-specific recommendations. The result is a system that transforms supply-chain updates into actionable project intelligence.
How we built it
To do so, we use an agentic AI framework over four layers designed for event extraction, project state updating, impact reasoning, and generating recommendations. The agentic layers ensure structured output, ready for systematic processing. Each agent can learn from specific examples and suggest optimizations for the specific site. All code is python based for fast execution, and the web-application is in HTML.
Challenges we ran into
One of the biggest challenges was realizing that construction already has planning systems. Simply building another dashboard or schedule viewer would not create meaningful value. We also encountered the problem of data availability. Real construction sites rarely have perfect, centralized information. Finally, we had to carefully determine where AI truly adds value. Many aspects of project management are deterministic and better solved through traditional software. We focused AI on information extraction, context synthesis, impact analysis, and recommendation generation.
Accomplishments that we're proud of
Developed a solution that goes beyond material tracking and focuses on project impact analysis.
What we learned
AI creates the most value when it synthesizes fragmented information and produces contextual understanding. Traditional software can store and retrieve data, but AI can help interpret changing project conditions and communicate impacts clearly. Finally, we learned that uncertainty is a first-class concern in construction. Any realistic AI system must reason under incomplete information and communicate confidence rather than pretending to know everything.
What's next for SiteMind
Our long-term vision is to evolve SiteMind from a project intelligence platform into a Construction Autopilot for site operations.
