AI success is all about metrics — adoption and engagement metrics, specifically. And original equipment manufacturer (OEM) CNH Industrial might have cracked the code on adoption with its AI assistant that is helping dealers repair farmers’ machines quickly and efficiently.

Today, CNH’s AI Tech Assistant has over a million global interactions with 30,000 active users, including 72% adoption across the OEM’s dealer network, after launching the service in January 2025. The OEM has made AI a key innovation priority for its $755 million agriculture R&D budget.

CNH started its AI journey focusing on dealers, given its nearly 20-year digital history with dealers and all the internal documents on its equipment, Chris Dempsey, VP of quality and product support at CNH, told AgNavigator. Dealers are using the AI Tech Assistant to find information quickly, including electrical schematics and operators’ manuals, he added.

This comes as dealers face two pressing challenges, a general technician shortage and a challenging market, Dempsey noted.

Farmers are struggling to stay on top of rising input costs, with U.S. tractor sales dropping by 21.6% and combines by 56.1% in May year-over-year, according to data from the Association of Equipment Manufacturers. Fewer tractor sales mean fewer tractors to service.

“Dealers’ profitability is largely driven by parts and services. So, the more efficient we can make our service business, the more efficient our dealers are going to be, and the health of our dealers in a challenged whole goods environment becomes much stronger,” Dempsey elaborated.

Titan Machinery becomes an AI Tech Assistant super user

One of CNH’s leading dealers, Titan Machinery – which operates over 90 locations across the globe – is a super user of AI Tech Assistant with over 78,148 interactions, or 7.8% of the total 1 million interactions.

Titan Machinery is using tools like CNH’s AI Tech Assistant to “do things that aren’t humanly possible,” combing through piles of data and surfacing insights for service technicians to solve problems, whether they are in the shop or in the field, Brad Meyers, director of service operations at Titan Machinery, told AgNavigator.

AI is not only providing insights on how to repair new tractor or combine models, but the technology is also filling a knowledge gap when it comes to older machine models, where insight might be hard to come by, Meyers explained.

“We have 40- and 50-year-old tractors on the farm we still use every day. But that knowledge of the technicians who used to work on that is retiring, and so where do we find tools to fill that gap?” he elaborated.

Despite the overall success, the adoption of AI Tech Assistant has not been uniform, Meyers admitted.

AI adoption “is more of an experiment in psychology than it is in technology,” with new and younger employees gravitating to AI tools and experienced employees using their tried-and-true methods, Meyers noted.

Also, Titan Machinery is exploring how AI can be pervasive throughout the service department, proactively suggesting solutions to open work orders, Meyers said.

“At any one time, we have 20,000 open work orders. … Nobody’s reading all of them every day. But if there’s a way to understand status of some of those and say, ‘Hey, I see you’re working on this unit with this diagnostic trouble code that isn’t solved yet, look you can use the AI Tech Assistant tool and might give us some advice,’” he elaborated.

Will AI ever get to a 100% accuracy?

Among its AI metrics, CNH reports a 90% accuracy with its assistant and as more data is fed into the system that percentage should move closer to 95%, Dempsey said. However, developing an AI tool is not just about accuracy, it’s about ensuring that technicians are overall more productive, he added.

“Getting to truly 100% accuracy, where AI does not need to engage with a human, or a human doesn’t need to be in the loop, I think we’re probably a ways away from that,” Dempsey said.

He added, “I want to see us make 100% of technicians more efficient — that’s really the goal. And certainly, I would love to see our tool evolve to be 100% accurate, but it really starts with how we can make our technicians more efficient.”