Enterprise logistics software has traditionally focused on tracking shipments, organizing data and reporting exceptions. GoComet is taking a more autonomous approach with Nova, an AI-native execution layer designed to carry out operational tasks across the freight lifecycle, from planning and procurement to documentation and payment.
GoComet introduced Nova at its Odyssey supply-chain leadership event in Singapore, where more than 150 logistics and supply-chain executives from companies including Shell, BHP, Bayer, L’Oréal and DP World gathered to discuss the next phase of enterprise AI.
The company’s central proposition is straightforward: logistics software should increasingly do the operational work rather than simply report what is happening.
Global merchandise trade reached approximately $25 trillion in 2025, according to UN Trade and Development. Yet many of the processes connecting orders, carriers, documents, bookings and payments still involve email, spreadsheets, phone calls and manual reconciliation.
That creates an execution gap. Supply-chain decisions can change quickly because of tariffs, capacity constraints, supplier delays or route disruptions, but implementing those decisions can require hundreds or thousands of follow-up actions.
Nova is designed to automate that coordination layer.
From Logistics Visibility to Autonomous Execution
GoComet says Nova builds on more than eight years of cross-border logistics data and context spanning more than 50 million shipments, 500 enterprises and over 70 countries.
The platform can monitor operational inboxes, identify document problems, create and resolve service tickets, compare contract and spot freight rates, and book shipments subject to defined cost controls. The company demonstrated a human approval step for decisions requiring additional oversight.
The architecture is important because logistics workflows are highly contextual. A freight decision can depend on a shipment, carrier, rate agreement, document, tracking status, invoice and organizational policy simultaneously.
Rather than treating each task as an isolated AI interaction, Nova is designed to retain that context across the workflow.
That distinction separates an agentic execution system from a conventional AI assistant. An assistant may recommend an action or summarize information. An agent can potentially retrieve the required information, make a permitted decision, update enterprise systems and follow through until the task is completed.
Humans Retain Decision Authority
GoComet’s model does not eliminate human involvement. Instead, it divides operational responsibilities between automated execution and human judgment.
Routine activities such as information retrieval, reconciliation, system updates and follow-ups can be delegated to AI agents. Humans remain responsible for higher-impact decisions, negotiations, complex exceptions and trade-offs.
That approach mirrors a broader enterprise AI trend toward human-in-the-loop agentic workflows, where autonomy is constrained by permissions, approval gates and business rules.
The challenge is particularly significant in logistics because an incorrect automated action can have physical and financial consequences. Booking the wrong carrier, approving an incorrect rate or mishandling a customs document can affect an entire shipment.
Nova’s execution model therefore depends not just on the underlying AI models but on the surrounding context, controls and enterprise integrations.
GoComet describes this longer-term vision as the “Invisible Hand of Global Trade”—software operating behind the scenes to coordinate routine activity while supply-chain professionals focus on decisions and exceptions.
Market Landscape
Enterprise AI is moving from copilots that summarize information toward agents capable of executing multi-step workflows.
Microsoft, Google, AWS, Salesforce and other enterprise technology providers are developing agent platforms that connect AI models with business systems and tools. In supply-chain technology, the same shift is emerging around procurement, logistics planning, inventory management and exception handling.
The logistics sector presents a particularly complex environment for agentic AI because workflows cross organizational boundaries. Carriers, suppliers, freight forwarders, customs providers and enterprises must exchange structured and unstructured information before a shipment can move.
For these systems, the competitive differentiator is therefore likely to extend beyond model intelligence. Access to operational context, enterprise integrations, workflow permissions, auditability and reliable execution will determine how much autonomy organizations can safely provide.
Top Insights
- GoComet’s Nova moves AI beyond logistics reporting by executing tasks across freight planning, documentation, procurement and payment workflows.
- The platform maintains shipment, rate, carrier, document and invoice context across multiple operational steps rather than treating tasks independently.
- Nova can automate routine coordination while retaining human approval for higher-impact decisions and exceptions.
- The launch reflects a broader enterprise shift from AI copilots toward agents that can interact with business systems and complete multi-step workflows.
- Logistics provides a significant test case for agentic AI because automated decisions can directly affect physical shipments, costs, schedules and business relationships.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
