Daimler Truck operates at the center of freight movement, manufacturing the vehicles that keep goods flowing across supply chains in North America and beyond. Its business spans trucks, buses, chassis, and powertrain systems, serving a complex industrial ecosystem where production, parts availability, and supplier coordination directly shape operational performance. In that environment, even small delays in planning and communication can ripple outward quickly, making speed, precision, and adaptability essential.

In this second part of a three-part CDO Magazine interview series, Edgar Gallo, Chief Data Officer at Daimler Truck North America, speaks with Susan Wilson of Alation, about how AI agents are beginning to address real manufacturing and aftermarket challenges. The conversation moves from theory to application, focusing on the practical ways agentic AI can reduce repetitive work, support planners, and help the business respond faster to changing demand.

In Part 1 of this series, Gallo spoke about how AI agents and metadata are reshaping manufacturing. The discussion focused on building trust, preparing the organization culturally, and using metadata as the foundation for agentic AI at scale.

From AI governance to grounded use cases

Gallo begins by acknowledging the scale and complexity of AI activity in a large enterprise. He notes that he does not have visibility into every AI use case across the company, particularly in an environment where many teams are already experimenting and innovating. What Daimler Truck has done, he explains, is first set up safety frameworks, establishing boundaries around responsible use.

That foundation allows the company to move from scattered experimentation toward clearer business use cases. Gallo describes working closely with a peer who leads the AI practice from a usability standpoint, while he focuses on the data and business side of the work. Together, those perspectives help translate operational pain points into practical agentic applications.

One of those opportunities emerges from the aftermarket business, where a leader raises a familiar problem: too much time is spent identifying the parts behind a plan, coordinating with suppliers, and handling all the surrounding tasks required before action can even begin. For Gallo, that signals a place where AI could remove friction without replacing human expertise.

Addressing the bullwhip effect

At the center of the discussion is the bullwhip effect, a familiar supply chain problem in which demand changes faster than planning and replenishment can keep pace. Gallo describes it in operational terms as a failure to catch shifts in demand quickly enough.

The challenge becomes more acute when a material planner is responsible for too many suppliers while also carrying a long list of repetitive tasks that must be completed before a response can be made. In that situation, the knowledge worker is not failing because of poor judgment, but because the volume of manual work gets in the way of that judgment.

Gallo sees AI agents as a way to change that equation. The objective is not to eliminate the planner’s role, but to make that role more effective by clearing away the mechanical work that slows it down.

A key idea in Gallo’s account is that many jobs combine two very different elements: decision-making based on knowledge and the execution of the tasks that follow from those decisions. He mentions that a comment from Nvidia CEO Jensen Huang helped crystallize that distinction for him, especially around the future of work and automation.

“If we can take any job and understand that it’s made of decision-making via knowledge and the task to execute the decision, and facilitate the task to happen on its own, you have more time to think and do value-added work.”

Teaching the agent how work actually happens

To make that vision practical, Gallo says the team works through multiple iterations to break the planner’s job into parts. The effort is not abstract. It is rooted in actual workflow.

“We went over several iterations until we were able to tell the agent, ‘This is a vertical piece… And then we went to the horizontal task — put it on an email, make an attachment, write the body of the email, identify who it goes to, and send the email.”

That decomposition matters because it lets the agent take over specific, structured work while leaving human judgment where it belongs. Once those vertical and horizontal tasks are absorbed by the system, the planner is freed to focus on the more strategic part of the role: discussing constraints with suppliers, documenting the plan, and evaluating the level of confidence around next steps.

Finding the right human partner for the pilot

Gallo makes clear that the success of an agentic use case depends not only on technology but also on the mindset of the person working alongside it.

The ideal participant, as he describes, is someone unafraid of retooling their job and understands that their talent is not threatened simply because the tasks around it may change. That distinction is central to his approach. The human contribution remains valuable precisely because it goes beyond mechanical repetition.

Moving from caution to confidence

As the use case matures, Gallo describes a moment when the team is ready to move past controlled experimentation. He recounts a meeting in which he pushes them to stop treating the system as something fragile.

“We had a call with them, and I said, we need to take the training wheels off. It’s time to pilot…” The team’s response surprised even him: “They not only said we’re ready, they said, but we don’t want to do it with one supplier; we have informed five; they’re ready, so let’s go with five.”

That moment reflects more than excitement. It signals trust. It shows that the individuals using the tool, the leaders overseeing the initiative, and the business partners involved all see enough value to move more quickly.

When asked what specifically gives the team the confidence to expand the pilot, Gallo points to a simple but powerful form of validation: the planner and the agent arrive at the same result.

That alignment matters because the planner is not validating the system by inspecting code or SQL syntax. Instead, he is challenging the business outcome. He uses his own process, filters his own data, applies his own thresholds, and compares that conclusion with the agent’s conclusion. When the two match, trust deepens.

Gallo explains that the planner’s reasoning process is layered and conditional. He may examine one column, then another, and then decide whether a third variable matters based on thresholds and experience. The agent’s usefulness lies in its ability to replicate that decision path closely enough to produce the same answer.

Reinforcing supplier relationships, not replacing them

For Gallo, the end goal of the initiative is not simply faster processing. It is stronger decision-making and better coordination with suppliers.

“We wanted to reinforce the relationship between their material planner and their vendors in the first place. That’s the focus now.”

That point reshapes the entire use case. The agent is not there to reduce human interaction but to create more room for meaningful interaction. Removing repetitive actions gives planners more time to engage suppliers, discuss plans, and work through uncertainty more thoughtfully.

Gallo even sees signs that this is already happening. The team’s decision to move from one supplier to five suggests that conversations are taking place and confidence is building not just internally, but across partner relationships as well.

CDO Magazine appreciates Edgar Gallo for sharing his insights with our global community.