While much of the market has moved past panic over software-as-a-service disruption, many enterprises still favor a custom, forward-deployed engineering mindset when approaching artificial intelligence. Yet, CJ Logistics charted a different course. Rather than building proprietary models from scratch, the company teamed up with its incumbent software vendor, Infios, to inject agentic AI directly into its core supply chain workflows.

Operating as a division of the major South Korean enterprise CJ Group, the North American branch of CJ Logistics anticipates closing out the year with approximately $1.1 billion to $1.2 billion in revenue. Its footprint spans the United States, Canada, and Mexico, with roughly 40 percent of business derived from warehousing solutions, another 40 percent originating from managed transportation, freight brokerage, and common carriage, and the rest dedicated to freight forwarding.

“We’ve never really been a low-cost provider. We’re typically a partnership-driven solver of problems in a customer supply chain.” — Brad Nuffer, COO

The enterprise oversees 25 million square feet of warehouse space distributed across 68 distinct facilities and manages $4 billion in annual freight, predominantly through managed transportation frameworks. When participating in common carriage, Nuffer emphasizes that the firm acts primarily as a strategic troubleshooter, embedding trucks within dedicated fleets to resolve operational bottlenecks near production plants. These high-capacity dedicated services frequently integrate tightly with local warehousing or freight forwarding infrastructure.

Core elements of the managed transportation division rely heavily on software from Infios—a platform CJ Logistics has utilized since 2010 when it operated under the MercuryGate banner. Supporting managed transport demands seamless interoperability with client warehouse management systems and occasionally manufacturing execution platforms, a flexibility where the Infios solution excels. It allows the consolidation of multiple client loads onto single trucks, supported by complex allocation and automated billing logic.

“We went through multiple versions of MercuryGate, grew with them, and then started working with their warehouse product,” Nuffer stated. The technology landscape shifted when Körber Supply Chain Solutions acquired MercuryGate, subsequently rebranding the entire entity to Infios.

By 2025, CJ Logistics integrated the Körber WMS architecture into its ecosystem. Although the firm runs multiple warehouse management systems, Infios remains a preferred tool due to its capability to quickly and cost-effectively launch multi-tenant facilities, enabling integrated inventory consolidation and shipping programs.

Although the acquisition was unplanned, Nuffer expressed enthusiasm for the outcome. “We lucked into a connection point between those two pieces of technology. We’re very excited about having that unified supply chain stack.”

Choosing Between LLMs and Dedicated Software Partners

Reflecting on the AI boom that swept logistics technology years prior, Nuffer noted the market confusion. “We saw some large competitors spending billions of dollars on AI investments. But what were they doing? It was very confusing as a buyer. No one knew how to really articulate the value.”

“We vetted all of them. And one of the core things that we always like to put front of mind when we make a decision on a technology investment is: Have we actually optimized the core operating system first before we bolt on something that might solve another problem?”

While evaluating alternative external AI providers, leadership weighed the friction of maintaining fragmented data sources and custom integrations. Ultimately, partnering with Infios offered a streamlined path. Because the Infios engineering team intimately understood their own workflow triggers and platform architecture, they could bake agentic intelligence directly into the core operating system, minimizing redundant clicks and maximizing decision-making speed.

The Infios and CJ Logistics AI Deployment

Rather than tackling complex, high-stakes tasks like carrier price negotiations, CJ Logistics targeted frontline inefficiencies—specifically, chasing down carriers that failed to transmit mandatory status updates. Despite widespread reliance on electronic data interchange (EDI) and visibility tools, drivers and dispatchers frequently neglect manual inputs, forcing human operators into an endless loop of unreturned phone calls and emails.

Automating this communications bottleneck yielded immediate insights into process design. Training an AI agent to execute follow-ups required designing intricate memory structures and rule-setting protocols. Establishing decision trees—such as determining whether to initiate contact via email or phone call, tailoring actions to specific carrier profiles, and setting follow-up frequencies—sparked valuable internal strategy sessions.

Within a couple of months, the team advanced from initial concept to a fully operational agent. The system automatically handles computer-generated voice calls, leveraging large language models to extract contextual reason codes for transit disruptions, such as highway congestion or extended warehouse dwell times.

With thousands of loads monitored daily, exception management remains critical. Clearer, faster data translates directly to superior customer service metrics. Furthermore, the conversational logic adapts dynamically; initial calls politely gather transit data, while non-responsive partners encounter increasingly firm automated follow-ups.

“While the person probably knows they’re talking with an AI agent, at least they know that if they give the agent good information, they’re not going to get another phone call,” Nuffer observed. Unresolved issues are seamlessly escalated to human operators.

Measurable Return on Investment

The financial payoff has been substantial. Previously, frontline personnel spent 20 to 30 percent of their shifts gathering routine location updates. Reclaiming those hours allows staff to redirect energy toward value-adding initiatives. As automated enforcement encourages carriers to comply with tracking standards, the operational environment becomes less exception-driven, driving down the long-term cost per load.

Pricing for the capability is usage-based, tied directly to token consumption via Infios. Because inbound calls require deeper contextual analysis than outbound pings, they consume more tokens. By balancing outreach frequencies against labor hours saved, CJ Logistics built a robust ROI model confirming the financial viability of the initiative.

Next on the roadmap is automated appointment scheduling, a notoriously non-uniform process that varies widely by retail receiver. Deploying tailored agents designed to navigate specific enterprise scheduling portals—such as those used by Walmart or Target—addresses another repetitive, labor-intensive workflow. Additional use cases include automating missing proof-of-delivery (POD) documentation alerts within the transportation management system.

Looking further ahead, decision-driving agents will assist optimization engines with routing and carrier negotiations. Warehouse operations represent the next frontier, where agents will eventually help manage inventory cutoffs, service level agreements, space utilization, and daily task prioritization.

Key Lessons from the Automation Journey

Reflecting on six months of intensive development, Nuffer noted that identifying business opportunities is simple compared to redesigning the underlying processes. Supply chain workflows were historically compromised by labor constraints; however, the introduction of continuous, scalable digital bandwidth calls for a total reassessment of traditional operational shortcuts.

For logistics providers anchored to core WMS and TMS architectures, collaborating closely with established software vendors offers the most pragmatic path forward. Ultimately, writing code for an agent is vastly quicker than aligning human stakeholders on how augmented workflows should function.