CHICAGO—For a consumer placing an online order, it’s a matter of a few clicks in a shopping portal. But the execution in a global supply chain requires multiple companies with multiple platforms to produce, ship, and deliver a product.

But often, the system doesn’t work as well as it should. Inefficiencies and exceptions to the plan make up the “swivel chair tax,” or the scramble to coordinate supply chain communication, as Ken Wood, executive vice president, product management at Descartes, explained.

Advancements in artificial intelligence will go a long way in eliminating those challenges, but such a solution “has to be curtailed and put inside workflows that you can trust, and it has to be observable,” Wood explained, opening the Descartes Innovation Forum here. The event this year brought together a record number of customers, speakers and sponsors for three days to explore innovation in supply chain and logistics technology.

And what better way to launch the gathering than a demonstration of the Descartes Agent Control Plane, designed to build trust in AI agents by coordinating them across applications and trading partners to execute multi-step workflows within defined guardrails, policies, and human oversight.

The demonstration showed how the technology manages a shipment—in this case, an armchair—across 11,000 miles, four companies, 15 Descartes products, and 28 steps.

“Even with connected systems, supply chain teams devote significant time to interpreting updates from emails, messages, and portals, rekeying information, and coordinating next steps to move orders forward,” Wood said. “The Agent Control Plane brings our logistics capabilities together so AI agents can take on more of that work across applications, conversations, and company boundaries; keep it moving as replies and approvals arrive; and act on events as they occur within permissions and approval points defined by customers.

“It’s an important step in our AI strategy, helping customers build on the Descartes solutions and AI platforms they already use and apply AI to the operational outcomes that matter, from reducing manual follow-up to keeping customer commitments on track.”

Built on the Descartes Global Logistics Network (Descartes GLN), the Agent Control Plane combines the operationally ready data and connectivity of the GLN with capabilities

from Descartes solutions to help AI agents coordinate work across applications and trading partners, and to make Descartes agents available within the AI platforms customers already use.

The GLN supplies the logistics context to connect changes in one part of a shipment’s journey with actions needed elsewhere, while the Agent Control Panel applies customer-defined permissions and approvals and maintains a traceable record of agent actions. Together, they help reduce manual handoffs, accelerate responses to changing conditions, and give teams more time to focus on exceptions and customer relationships.

From factory floor to living room

In a demonstration using 15 actual Descartes solutions and team members to represent the shipper, the freight forwarder, and the freight broker, the Agent Control Plane coordinated the various shipping handoffs, managed complications, and offered alternative solutions—doing in seconds what might otherwise require hours of human back-and-forth communication.

“It’s autonomy in action, but the agent isn’t doing everything; it's only doing the things that you allow it to do,” Wood said, as the virtual demonstration reached the final mile dispatch stage. “There's so much packed into this—it's crazy to those of you here that live in this space. But the truck is on the way. [The consumer] doesn't know anything about that. All he knows is his chair is still planning to arrive on time.”

The demonstration concluded with the sound of a doorbell.

“Our products have learned to work together in a way that they never could before, and we're very excited about that,” Wood concluded, noting that the Agent Control Plane’sdevelopment began in February. “This is just the beginning.”