Apex

By the time the disruption is visible, it is already too late.

A global logistics network generates thousands of active shipments a day across disconnected tools and manually updated spreadsheets: every carrier exposes different formats and cadences, weather and port data live in incompatible silos, and the average window to react to a disruption is a few hours, often already too late. The problem is not collecting the data, it is making it actionable in real time without drowning operators in noise.

72-hour risk forecasting

An ML model blends weather data, port congestion indices, and carrier reliability history into a single per-shipment risk score, recomputed every few minutes across a 72-hour horizon.

Unified data lake

A Kafka architecture with TimescaleDB normalises feeds from dozens of carriers and authorities into a single queryable schema, with fault tolerance and controlled backpressure that eliminate silos.

God-mode 3D globe view

A WebGL renderer displays thousands of active routes with animated arcs encoding risk severity, volume, and direction at 60 fps on standard hardware, the complexity of the whole network on a single screen.

Calibrated proactive alerts

A tiered alert engine fires only above calibrated confidence thresholds, drastically cutting alarm noise and focusing operator attention on the risks that genuinely matter.

Aggregate

Real-time feeds from carriers, ports, and weather services flow through Kafka into the data lake, normalised into a single time-series schema.

Score

The ML layer scores every active shipment every few minutes, producing a 72-hour risk score with rationale.

Visualise

The 3D globe makes the whole network navigable, with routes and risk encoded by colour, thickness, and motion.

Preempt

Calibrated alerts warn teams hours before a disruption turns critical, turning reaction into prevention.

Standards and certifications

Security management aligned to ISO 27001 and operational controls compliant with SOC 2 Type II, with a traced release pipeline and separation between development and production environments.

Data sovereignty

In on-premise and isolated modes shipment data and carrier contracts do not leave the customer perimeter; in cloud, region and data residency are configurable to respect commercial and regulatory constraints.

Access and audit

RBAC to separate operators, analysts and administrators; every alert, threshold override and data access is recorded in an immutable audit trail for review and compliance.

Encryption

AES-256 encryption at rest for the data lake and TLS 1.3 for all inbound feeds and outbound APIs; carrier credentials are stored in an encrypted vault with key rotation.

How many carriers can we connect?

The Kafka architecture is designed for dozens of simultaneous heterogeneous feeds; controlled backpressure absorbs different cadences and volumes without degrading scoring. New carriers are added via connectors without stopping the platform.

How do you avoid alert fatigue?

The notification engine is tiered and fires only above calibrated confidence thresholds: in deployments it cuts alert noise by 74%, focusing operator attention on the risks that matter.

Is the risk explainable or a black box?

Every risk score comes with the rationale of the factors driving it, weather, port congestion, carrier reliability, so operators understand why a shipment is at risk and can act accordingly.

Can we integrate it with our ERP?

Yes. A GraphQL API exposes risks, states and forecasts to downstream systems, so Apex feeds existing ERP, TMS or dashboards without forcing operators to switch tools.