AI Can Analyze the Data, but Humans Still Own the Outcome

Artificial intelligence has become the loudest conversation in trucking, promising faster compliance checks, deeper video analytics, and instant answers to complex regulatory questions. But according to Rustin Keller, President and CEO of J. J. Keller & Associates, Inc.

, the technology’s growing footprint hasn’t changed who’s ultimately accountable when something goes wrong. “AI is going to accelerate business, not just in high tech places where it can be awe inspiring, but also in the back office,” Keller said. “It gives employees the ability to quickly do analysis of a lot of data. It’s amazing.”

Speed and insight, though, don’t remove responsibility. “At the end of the day, humans are still responsible for outcomes,” Keller said. “Humans still have to run the operation, site, or fleet safely. You’re responsible for incidents and the programs you implement to make things safer. When an incident occurs, you don’t get to say ‘AI is to blame.’”

Rustin Keller [Credit: J. J. Keller & Associates, Inc. ] Fleets today have access to a volume of operational data that would have been unimaginable a few years ago. The value of that visibility is immense, but it doesn’t solve every problem. “You now have footage of the drivers, yards, warehouses, and all aspects of the operation,” Keller said.

“You now can get more data than ever on all accidents from a given period and you can analyze root causes.” The harder problem, in his view, starts after the data is in hand. “You still need to know what to do next,” Keller said. “Getting better information is awesome, but now what?” AI can generate options, but it can’t make the call.

“You still own it and you still have to decide what’s appropriate. AI can’t understand your team and your operation’s nuances. You have to take the next step forward.” The gap between data and action is where experts bring the most value. “That’s what we do at J. J. Keller,” Keller said.

“We make great safety and compliance products and platforms, but we’re an advisory company first.” J. J. Keller’s identity as an advisory firm is baked into the company’s origin story, and Keller traces it back to his grandfather. The company was founded in 1953 by John J. “Jack” Keller in Neenah, Wisconsin.

“My grandfather started the company as a one-man consulting firm to help motor carriers navigate complex new interstate trucking regulations under the ICC and DOT,” Keller said. “Our first product was advising customers, and that’s still our roots.”

That advisory pedigree shows up in engagement numbers today, and Keller says demand for direct human expertise has climbed despite the advent of new technologies. “We do more than 350 webcasts and seminars every year,” Keller said. “Some people might think AI will slow that down, but we’re seeing all time records on attendance.

More people are seeking our seminars than ever before.” According to Keller, what’s driving that appetite is not AI adoption itself but what he considers the industry’s real adversary. “We always say that our number one competitor is complacency,” he said. “Doing nothing is number one. Companies often decide that what they have is just fine.”

What’s shifting, he argued, is fleets’ willingness to tolerate that complacency. “What’s happening now, though, is that people are getting more educated and now are seeking more information from experts when they run into the difficult issues,” Keller said. “More people can see the pitfalls and the legal cases, and they know they have to do something.”

Keller has also noticed a change in tone from customers who once pushed back on regulatory requirements they didn’t want to accept. “Customers used to argue with our consultants on the little nuances of regulations,” he said.

“They would insist they didn’t need CDLs or hazmat certifications for their operations when they most certainly did, but that kind of attitude is largely going away.” Basic regulatory facts are easier to find than ever, he said, but that hasn’t eliminated the need for guidance. “AI can’t sit down with you and talk about improvement plans.” Even as J. J.

Keller builds its own AI capabilities, there are limits to what technology can accomplish. “Our team is building AI tools, trained off of our own data,” Keller said. “But AI can only do so much. It’s making its best estimate. Ultimately, it’s still up to our consultants to walk through real actionable steps.”

With so many vendors marketing faster, AI-driven answers, Keller offered a simple standard for separating substance from hype: can the tool’s output hold up under scrutiny? “Can your software defend an approach you decided to take or processes you implemented?” Keller asked. “Can it explain why you made a choice or exercised professional judgment?

You still need experts to explain the options you had and what course you took and why.” That standard, he said, also draws a line between using AI to gather information and using it to make decisions that carry legal or safety weight.