CargoAi today announced the launch of AI Predictive Tracking, a new capability designed to help air cargo stakeholders anticipate operational risks and shipment delays before they materialise.
The solution is available both within CargoMART and as an add-on to the CargoCONNECT Track & Trace API.
The release reflects a growing operational challenge in air cargo: while traditional tracking tools provide visibility once milestones are reported, they often leave limited time to act when a shipment is at risk.
CargoAi’s Predictive Tracking introduces an additional layer of intelligence by forecasting upcoming shipment events and triggering early risk alerts.
AI Predictive Tracking uses machine learning models trained on millions of historical shipments, combined with live airline flight updates, to predict the expected timing of each key milestone in an air cargo journey.
These include documentation submission (FWB), acceptance (RCS), manifesting (MAN), departure (DEP), arrival (ARR), freight availability (NFD) and final delivery (DLV).
Instead of relying solely on reported events, the system generates probability-based predictions, including median (P50) and conservative (P90) estimates.
These predictions are continuously refreshed as new information is received, enabling operational teams to detect risk patterns earlier in the process.
Those prediction values can be used as is by our API users, but are also translated into comprehensible signals.
Turning predictive signals into operational action
The predictive layer is designed to support concrete operational decisions across the air cargo ecosystem:
- Airlines can identify shipments that have not reached RCS or MAN before cutoff and are therefore at risk of missing their planned flight. Alerts can be triggered to stations or GSAs to intervene, release blocked capacity, or prioritise high-risk shipments. The data can also be used to benchmark station performance and identify recurring bottlenecks.
- Freight forwarders gain early visibility on shipments flagged as “At Risk,” allowing them to act on missing documentation, coordinate pickups, or proactively inform customers. Predicted cargo availability (NFD) can be shared downstream to improve planning and customer communication.
- Ground handling agents and system integrators can use conservative P90 predictions to prioritise acceptance, automate pre-alert management, and feed predictive risk levels into internal dashboards or SLA monitoring tools, reducing the need for manual checks.
Example scenario
On a CDG–SIN flight scheduled to depart at 18:00, the system may detect at 10:00 that the FWB has not yet been received, whereas historical and live data showed it should have happened in 90% of the cases already.
An alert is generated indicating a high probability of missing the planned flight. Operations teams are notified and can talk to each other to confirm actions.
The predictive engine combines:
- Historical performance data by airline, route and product type
- Live flight schedules and operational events
- Standardised milestone structures aligned with CargoIMP and IATA ONE Record
Each milestone is enriched with predicted timestamps and confidence levels. An alerts object highlights current risk levels—low, medium or high—along with contextual messages to support decision-making.
The solution is fully backward-compatible, and existing Track & Trace API integrations remain unchanged.
AI Predictive Tracking is available across both CargoAi’s platform and API ecosystem, supporting a range of operational and technical environments. The solution can be activated using existing CargoAi connectivity, with alerts automatically refreshed whenever milestones are updated or flights are rescheduled.
Within CargoMART, AI Predictive Tracking is embedded into enterprise-grade operational workflows:
- CargoBridge Integration connects CargoMART directly to a customer’s TMS or ERP, consolidating shipment data in a single environment and eliminating duplicate data entry.
- Predictive AWB Tracking (AI-powered) delivers early delay alerts and proactive status updates across shipments, enabling operational teams to identify risk and intervene before cutoffs or missed flights.
Via CargoCONNECT, AI Predictive Tracking is offered as an add-on to the Track & Trace API:
- Predictive milestones are included directly in API responses, with P50 and P90 estimated event times.
- A structured alerts object indicates current risk levels and provides contextual messages that can be integrated into internal systems, dashboards or automated workflows.
- The feature is fully backward-compatible, allowing existing integrations to remain unchanged.
This dual availability enables organisations to apply predictive intelligence either through user-facing operational tools or directly within their own systems and applications.