Adds Critical Data to Global Logistics Network and Advances AI-Powered Fleet Safety and Performance Management Capabilities

WATERLOO, Ontario and ATLANTA, Georgia, April 23, 2026 (GLOBE NEWSWIRE) — Descartes Systems Group (Nasdaq:DSGX) (TSX:DSG), the global leader in uniting logistics-intensive businesses in commerce, announced that it has acquired Idelic, a leading provider of AI-powered driver safety and performance management solutions.

Idelic’s safety intelligence platform unifies day-to-day safety management activities such as training, monitoring, reporting and coaching into a single solution. With an AI analytics workflow designed to measurably impact driver behavior, fleets can proactively identify and reduce driving risk by leveraging the platform’s unique dataset of more than 40 billion miles of telemetry and over 400,000 accident records. The company collects real-time, highly detailed event-level data through a connected network of more than 80 telematics, risk management, and regulatory system integrations. Built on years of machine learning applied to predictive accident models across more than one hundred and fifty fleets, Idelic’s AI capabilities are field-proven in predicting driver risk and optimizing safety training interventions.

“Productivity and safety are equally critical for fleet operators,” said James Wee, General Manager of Fleet Management at Descartes. “This acquisition adds critical data to our Global Logistics Network (GLN) and enhances Descartes’ final-mile footprint by adding highly advanced fleet safety capabilities and deep domain expertise. Idelic’s AI-powered predictive safety intelligence functionality, when combined with Descartes’ industry leading routing planning and execution technology, enables us to deliver a complete and cutting-edge fleet performance management solution that uniquely incorporates driver behavior and safety signals into our robust operational data set.”

“The need for trusted, real-time fleet and operational data is becoming increasingly critical as customers advance their AI strategies,” said Edward J. Ryan, Descartes’ CEO. “By combining Idelic’s predictive safety intelligence and unique, critical data with Descartes’ GLN, we’re strengthening the data foundation that powers smarter fleet performance decisions and safer, more efficient operations.”

Idelic is headquartered in Pittsburgh, Pennsylvania. Descartes acquired Idelic for up-front consideration of approximately US $28 million satisfied with cash on hand, plus potential performance-based consideration. The maximum amount payable under the all-cash performance-based earn-out is US $12 million, based on the combined business achieving revenue-based targets in each of the first two years post-acquisition. Any earn-out is expected to be paid in fiscal 2028 and fiscal 2029.

About Descartes

Descartes powers more responsive, efficient, secure and sustainable international and domestic supply chains by uniting logistics-intensive businesses on its Global Logistics Network (GLN). Shippers, carriers, and logistics service providers connect and collaborate on the GLN leveraging technology, data and AI to manage last mile deliveries, domestic and international shipments, transportation rating and payment, global trade research, customs compliance and a variety of regulatory processes. Learn more about Descartes at www.descartes.com

The post Descartes Adds Predictive Fleet Safety Intelligence with Idelic Acquisition appeared first on Logistics Viewpoints.

Supply chain visibility has improved dramatically. But seeing a problem is not the same as deciding what to do about it. That is where many AI initiatives still fall short.

A supply chain manager no longer has to wait for a phone call to learn that a shipment is late. In many cases, the system already knows. The carrier feed has updated. The ETA has changed. The control tower has flagged the exception. An alert is sitting in the dashboard.

That is progress.

But it is not the same as resolution.

The harder question comes next: what should happen now?

A delayed shipment may be a minor inconvenience. It may also create a production shutdown, a missed customer commitment, or an expensive expediting decision. The system can show the delay. It may even recommend a response. But unless it understands the full operating context, the recommendation can be incomplete.

That is the gap between visibility and decision-making.

Visibility Solved the First Problem

Over the past decade, companies have invested heavily in supply chain visibility. Control towers, real-time tracking, event platforms, and data integration tools have made it easier to see what is happening across transportation, warehousing, suppliers, carriers, and customers.

That was necessary. For years, supply chain teams operated with too much delay and too little shared information. They learned about problems after the fact. They chased updates manually. They managed exceptions through email, spreadsheets, and phone calls.

Visibility improved that picture.

But visibility answers only the first question: what is happening?

It does not automatically answer the more valuable question: what should we do?

The Late Shipment Problem

Consider a late inbound shipment of components headed to a manufacturing site.

The visibility platform flags the delay. The AI model may suggest expediting the shipment by air. On the surface, that seems reasonable. The shipment is late. Air is faster. The recommendation appears logical.

But the real decision depends on context.

Is the component needed for tomorrow’s production run, or next week’s? Is there substitute inventory at another site? Is the customer order attached to this component strategic or routine? What is the cost of expediting relative to the margin on the order? Will pulling air capacity for this shipment create a problem somewhere else?

Without that context, the system is not really making a decision. It is responding to an event.

That distinction matters.

More Alerts Are Not Better Decisions

AI can increase the speed and volume of detection. It can identify more anomalies, generate more alerts, and surface more possible actions.

But most supply chain teams do not need more alerts. They need better prioritization.

A one-day delay on a low-value replenishment shipment may not matter. A six-hour delay on a critical component may matter a great deal. A missed delivery window for a strategic customer may require immediate escalation.

The value is not in seeing every exception. The value is in knowing which exceptions deserve action.

That requires decision logic.

Decision Logic Is the Missing Layer

Decision logic is the operating structure that turns a signal into a response. It defines service priorities, cost thresholds, inventory rules, customer commitments, capacity constraints, and escalation paths.

Most companies have this logic, but it is scattered.

Some of it lives in planning systems. Some sits inside transportation workflows. Some is buried in spreadsheets. Some exists only in the judgment of experienced planners.

AI cannot reliably automate decisions when the decision rules are fragmented or informal.

That is why many systems remain advisory. They can tell the organization what may be wrong, but people still have to decide what matters, determine the right action, and push execution across functions.

The Real Constraint Is Execution Authority

A recommendation is not an outcome.

If the AI system recommends expediting a shipment, something still has to happen. Capacity must be secured. Cost must be approved. Inventory plans may need to change. The customer may need to be notified. The production schedule may need to be adjusted.

If those steps are not connected, the system has not changed the operating model. It has only added a smarter alert.

This is where many AI pilots struggle. The model performs well in a controlled setting. It identifies the problem. It proposes a response. But in live operations, the recommendation runs into unclear decision rights, incomplete data, or manual workflows.

The issue is not always the model. Often, the issue is that the organization has not defined who or what has authority to act.

From Control Tower to Control System

The next phase of supply chain technology should not be measured by dashboard quality alone. It should be measured by whether systems can help execute better decisions.

That means moving from control towers toward control systems.

A control tower provides visibility. A control system connects visibility to context, decision logic, workflow, and feedback.

It knows which exceptions matter. It knows what options are available. It knows which decisions can be automated and which require human review. It records the outcome and improves over time.

That is a higher bar than visibility. It is also where supply chain AI becomes operationally useful.

The Late Shipment, Revisited

In the late component example, the issue is not simply whether the shipment is delayed. The issue is whether the delay changes production, customer service, inventory allocation, or cost exposure.

Until the system can connect those consequences, it is still reporting the problem rather than managing it.

That is the practical test for supply chain AI. Can it connect a signal to the business consequence? Can it recommend the right action? Can it route that action to the right person or system? Can it learn from the result?

If not, it is still a visibility layer with better language around it.

Humans Still Matter

None of this means removing people from supply chain management.

The point is the opposite.

Human judgment should be focused where it matters most: ambiguous exceptions, high-value trade-offs, customer-sensitive decisions, and situations where the consequences of error are significant.

Routine, low-risk decisions should be handled with more automation. Complex decisions should be escalated with better context.

That is the right division of labor between people and AI.

The Practical Path Forward

Companies do not need to solve this across the entire supply chain at once. They should start with the most common and costly decision points: late inbound shipments, inventory allocation conflicts, carrier exceptions, production constraints, supplier delays, and customer service failures.

For each one, the questions are straightforward:

What data is required? What constraints matter? What actions are allowed? What can be automated? What requires approval? Who owns the decision?

Those questions are not glamorous. But they are the foundation of useful supply chain AI.

Visibility was necessary. It was never sufficient.

The next competitive advantage will not come from seeing more. It will come from deciding better, faster, and with enough context to act.

The post Visibility Isn’t Decision-Making in Supply Chain AI appeared first on Logistics Viewpoints.

Discover Freightos Enterprise

The post Transpac ocean rates continue to edge up on Iran pressure – April 28, 2026 Update appeared first on Freightos.