AI adds new potential to truck specification

Randy Bell, associate director of digital solutions at International Motors. (American Trucking Associations)

PITTSBURGH — Randy Bell, associate director of digital solutions at International Motors, contrasted his big-tech background with the current state of truck specification tools during a discussion about artificial intelligence’s role in vehicle spec’ing.

“Historically, configuration systems have focused on one primary question: Can this truck be built?” Bell said. “But a buildable configuration isn’t necessarily the best configuration.”

On Sept. 23, Bell and other panelists explored that distinction, along with the promise and pitfalls of AI, during The AI Summit, held in conjunction with American Trucking Associations’ Technology & Maintenance Council Fall Meeting.

Bell outlined what he called a three-step “AI maturity curve” that included conversational AI assistants that help users navigate complexity; AI advisers that reason across fleet-specific data; and intelligent partners that learn from real-world outcomes across many fleets.

“The future isn’t AI spec’ing trucks,” he concluded. “It’s AI helping fleets and their OEM partners spec trucks better.”

Kristopher Ptasznik, Cummins’ fuel economy and powertrain TCO consultancy leader, cautioned that realizing that future will require learning hard lessons. He described a multiyear effort to develop an AI specification tool that ultimately was shelved.

“We worked with a third party for roughly two years, and at the end of two years, the tool wasn’t good enough,” Ptasznik said. “It didn’t provide the answers that we were looking for, at least not consistently enough.”

As one example, he said the model confidently recommended a higher road speed for a heavy-haul fleet based on data that failed to distinguish the operation from a typical linehaul fleet.

Image

Kristopher Ptasznik, Cummins’ fuel economy and powertrain TCO consultancy leader, says AI may not always make the best decision, "but it's always confident in its decsion." (American Trucking Associations)

“In no world will [driving faster] ever happen” for fuel economy in that case, Ptasznik said. “AI, when it doesn’t have those data points to pull in, it’s not making a good decision. But it’s always confident in its decision.”

On the OEM side, AI also is reshaping dealer tools and sales roles. Carson Howald, SpecPro operations manager at Daimler Truck North America, said expectations shaped by e-commerce are colliding with legacy specification systems.

Image

Expectations shaped by e-commerce are colliding with legacy specification systems, says Carson Howald of Daimler Truck North America. (American Trucking Associations)

“The configuration tool has to have improved onboarding,” he said, noting that proficiency can take years even as dealers increasingly hire people who “were selling refrigerators last week.” AI, he argued, must make the tools easier to use through proactive recommendations and better visualization.

But Howald stressed that AI will not eliminate the need for human relationships. As the tools handle more compatibility checks and data analysis, “the value proposition then begins to shift,” he said, toward spending time in the field and “really having a deep understanding of what our customers need and what they want.”

Image

“If you want to handle your data in a private way, you have to understand how to construct your system,” DG Technologies President Mark Zachos says. (American Trucking Associations)

DG Technologies President Mark Zachos warned that as AI agents become more capable, they also become more vulnerable to exploitation. He cited prompt-injection attacks and the risk that AI could be quietly biased toward certain brands or components.

“Data privacy, we’ve heard about that,” he said. “If you want to handle your data in a private way, you have to understand how to construct your system.”

Panelists agreed that strict guardrails, continuous testing and human oversight remain nonnegotiable. Bell said AI must operate within rules and values, with experts able to pull the plug whenever the technology behaves in ways that are “inconsistent with the way we set it up.”