From AI hype to real-world impact: How AI agents are transforming logistics

AI agents promise to transform transportation and logistics. But where should businesses start, how can they measure the value and what does it take to deploy AI safely?

At the recent FreightWaves Enterprise Fleet Summit, Megan Liss, Director of AI go-to-market strategy, joined FreightWaves' Thomas Wasson to discuss how Trimble is bringing agentic AI into real-world logistics operations.

Drawing on Trimble's own experience, Megan shared practical use cases, insights into enterprise-grade security and her predictions for the future of AI-powered logistics.

Walk a mile in their shoes

Before bringing Trimble Arc Agent to market, Trimble spent two and a half years building and using AI agents internally to automate repetitive workflows.

"We thought that it was really important that we walk a mile in our customers' shoes," explains Megan.

Working alongside its own AI-powered "digital colleagues" gave Trimble firsthand experience of the opportunities and challenges involved in deploying agentic AI, helping the company guide customers on where and how to use the technology.

For now, Megan sees AI primarily as a way to augment employees rather than replace them, taking on tedious tasks so people can focus on higher-value work.

Start with low-hanging fruit

For logistics companies looking to get started, Megan recommends focusing on repetitive, time-consuming tasks that create operational bottlenecks, particularly in dispatching and customer support.

Key opportunities include:

• Order entry and contract intake: Extracting information from emails and bills of landing, cross-referencing details and entering data directly into a transportation management system (TMS).
• Dispatch and customer support: Automating routine administrative tasks and customer communications.
• Track and trace: Streamlining tracking updates, dwell time and detention-related workflows.

The potential productivity gains are significant. Megan estimates that a dispatcher or back-office employee handling 15–25 loads per day could manage 50–80 with AI agent support, helping reduce the cost per load without adding headcount.

However, she emphasizes the importance of testing use cases and measuring results before scaling automation or making decisions about workforce requirements.

Building trust

As AI agents take on more operational responsibilities, security, transparency and control become essential.

Rather than building its own large language model, Trimble developed the Arc Safety Harness, adding enterprise-grade safeguards to existing AI models. These include testing and evaluating skills before deployment, along with audit logs that show exactly what work an agent has performed.

Megan also emphasizes the importance of keeping a human in the loop, with agents making recommendations that employees can review and approve.

Providing secure AI tools can also help businesses address the growing challenge of "shadow AI", where employees use personal AI accounts for work-related tasks, potentially exposing sensitive company information.

One platform, multiple AI skills

Another challenge for businesses is managing multiple AI agents across different workflows.

Rather than requiring companies to deploy separate agents for every task, Trimble has brought these capabilities together within Arc Agent.

Users can create skills through natural-language instructions, tailoring workflows to their specific processes and applications. Skills can also be shared with colleagues, while more complex tasks can be coordinated through automated orchestration.

This unified approach makes it easier to manage and scale AI adoption without introducing a growing collection of disconnected tools.

Looking ahead

Looking to the future, Megan predicts a shift towards management by exception, where AI agents handle routine logistics activities while employees focus on oversight, problem-solving and higher-value decisions.

She also anticipates more agent-to-agent interactions, enabling AI systems to coordinate and negotiate directly with one another across supply chain networks.

While this transition will take time, the potential is significant: fewer manual processes, more seamless collaboration and greater efficiency across complex logistics operations.

For businesses looking to get started, the message is clear. Focus on practical use cases, measure the value and put the right safeguards in place. That's how AI moves from hype to real-world impact.

Want to learn more?

Watch the full FreightWaves interview to hear Megan Liss discuss how agentic AI is changing logistics operations.