Sensing
The Sensing stage collects GPS pings, carrier API status events, and active weather alerts into a single stream. Each signal is scored against the shipment’s estimated transit time so that only events crossing the configured severity threshold — fifteen percent by default — generate an in-app alert. The result is a continuous picture of the lane without a dispatcher asking for it. The pipeline collapses three noisy feeds into one ranked DisruptionEvent the routing engine can act on, so the alert the moment it crosses the threshold arrives as a scored event, not a notification.
Negotiating
The Negotiating stage is where AI supply chain disruption response earns its name. For every new DisruptionEvent the routing engine receives the shipment context and returns a single scored reroute recommendation — a primary route plus the two next-best alternatives, with cost and SLA tradeoffs side by side. Pricing is sourced from current spot-market rates so the recommendation is a real call, not a placeholder. The on-call planner either accepts the recommendation from the dashboard or overrides it with a reason that is fed back into the engine, closing the loop between detection and approval on the same lane.
Self-healing
The Self-healing stage is what makes the next disruption quieter than the last. Every accepted reroute is logged as a rerouting event, and that outcome updates the lane’s confidence baseline, the corridor weather model, and per-carrier reliability scores in the background. The threshold itself adapts to the fleet’s baseline ETAs, so a slow corridor and a fast one stop sharing the same noise floor. The platform does not stop being autonomous between disruptions — it gets more accurate.
See it on a real lane
Both pricing tiers run the same autonomous disruption pipeline on your real fleet — pick the depth and keep the engine.