ZENITH

One shared, AI-powered intelligence platform that predicts delays, orchestrates decisions, and transforms construction supply chains

Comment

ZENITH – AI-Powered Construction Supply Chain Intelligence

Inspiration

Construction projects don't usually fail because materials don't exist—they fail because nobody has a complete picture of where those materials are, whether they're delayed, or who needs to act.

During our research, we discovered that construction loses over $1.6 trillion annually due to fragmented supply chains, poor visibility, communication gaps, and reactive decision-making. Existing software manages only isolated parts of the workflow—procurement, scheduling, BIM, or project management—but none provides a unified intelligence layer across the entire material lifecycle.

We wanted to build the missing layer: an AI-powered platform that predicts delays before they happen, connects every stakeholder through one shared source of truth, and recommends the best mitigation actions before projects are affected.

That vision became ZENITH.

What We Built

ZENITH is an AI-powered construction supply chain intelligence platform that tracks every material from procurement to installation while continuously analyzing project health.

Instead of simply recording events, ZENITH transforms operational data into actionable intelligence.

Our platform provides:

  • Predictive delay detection
  • AI-generated mitigation recommendations
  • Confidence-based ETA forecasting
  • Cross-organization material visibility
  • Vendor reliability scoring
  • Approval bottleneck detection
  • Regulatory impact monitoring
  • Executive project health dashboards
  • Human-in-the-loop decision support

Different stakeholders—including owners, procurement teams, vendors, logistics teams, and finance—work on role-specific dashboards while sharing one synchronized data model.

How We Built It

We designed ZENITH as a modular, event-driven architecture focused on scalability and explainable AI.

Frontend

  • React
  • TypeScript
  • Tailwind CSS

Backend

  • FastAPI
  • WebSockets
  • PostgreSQL
  • Redis Pub/Sub

AI Layer

We developed multiple specialized AI agents responsible for tasks such as:

  • Delay prediction
  • Risk explanation
  • Mitigation planning
  • Vendor profiling
  • Root-cause analysis
  • Approval intelligence
  • Regulatory impact assessment
  • Executive health narration

Rather than allowing autonomous execution, every recommendation is reviewed through a Human-in-the-Loop workflow, ensuring transparency and trust.

Research Process

Our solution was built after studying:

  • Construction supply chain workflows
  • Industry reports from McKinsey, Bent Flyvbjerg, and government publications
  • Existing construction software ecosystems
  • Procurement approval processes
  • Material lifecycle management
  • Delay propagation across multiple supplier tiers

We compared leading platforms and identified a common limitation: each solves only one segment of the construction lifecycle.

This research shaped the core architecture of ZENITH.

Challenges

Building ZENITH involved several challenges.

The biggest challenge was modeling a construction supply chain that spans multiple independent organizations while maintaining one consistent source of truth.

Another challenge was designing AI that remains useful without taking autonomous actions. We focused on explainable recommendations, confidence scoring, and decision support instead of automated execution.

Creating scalable event-driven workflows, keeping data synchronized across stakeholders, and ensuring every AI output remained transparent were also key engineering challenges.

What We Learned

This project taught us that successful AI systems are not just prediction engines—they are coordination platforms.

We learned how important explainability, human oversight, reliable system architecture, and stakeholder-specific experiences are when building enterprise AI products.

We also gained practical experience in:

  • Event-driven backend architecture
  • AI orchestration
  • Supply chain modeling
  • System design
  • Product thinking
  • User-centered interface design

Future Roadmap

ZENITH is designed as a scalable platform.

Future versions will include:

  • ERP integrations
  • BIM and IFC connectivity
  • IoT-enabled material tracking
  • Digital Twin integration
  • Computer vision for site verification
  • Predictive procurement optimization
  • Multi-region regulatory intelligence
  • Enterprise deployment across large construction portfolios

Impact

Our goal is simple:

Transform construction supply chains from reactive tracking systems into intelligent decision-making platforms.

By predicting delays before they occur, improving collaboration across organizations, and providing explainable AI recommendations, ZENITH helps construction teams reduce delays, lower costs, improve transparency, and deliver projects more reliably.

Built With

Share this project:

Updates

Submission history