pixone3d - Fotolia
SAN FRANCISCO -- John Deere, the brand name of Deere & Company, brings to mind green tractors in a golden field. It elicits thoughts of the earth, of planting and growing, of hard labor. Yet, for this classic American corporation, those thoughts are only part of the picture.
John Deere has manufactured and sold agricultural machinery and equipment for more than 180 years. It's one of the biggest farm machinery manufacturers in the world. Over the last few years the multibillion-dollar company has made a significant progress on its AI journey, to develop AI-driven technology and embed it into its machines.
Not necessarily new
But developing and using advanced technologies isn't new to John Deere, said Julian Sanchez, director of precision agriculture at John Deere, in an interview at the AI Summit conference Sept. 26.
For some 25 years, the company has put GPS capabilities into its tractors and other machines, enabling farmers to track their work. John Deere has also built self-driving machines for more than two decades.
"We're a company that has very, very quickly reinvented itself from a hardware manufacturer to a developer" of software and AI, Sanchez said.
Despite the company's long history of developing machinery and technology, making that major push to create advanced software didn't happen quickly.
Julian SanchezDirector of precision agriculture, John Deere
"It really started with recognizing that we are rapidly becoming a software company" more than a decade ago, Sanchez said. John Deere began recruiting heavily, looking for talent from universities and research programs.
To advance its AI journey, company focused heavily on developing software teams and creating a software culture.
For example, John Deere maintains an internal list of which languages employees can speak, Sanchez said. They began adding programming languages to that list and hiring large numbers of software developers at a time.
Marrying hardware with software
It was difficult to merge the equipment with the software, Sanchez said.
"We've been working very hard the last decade to marry those two," he said.
On the hardware side, the company rolled out significant changes to its machinery about once a year. Yet, with software, changes can be introduced much faster, sometimes as often as weekly. Putting out significant key updates to bring new features to older pieces of hardware wasn't easy, Sanchez said.
Still, he added, the last two model years of harvesting combine machines have received significant feature updates, adding new capabilities, without having to change any hardware.
The work appears to have paid off. A number of John Deere machinery can automatically perform farming actions with little to no real-time human input. For example, Sanchez said, the company makes harvesters armed with video cameras.
With computer vision and machine learning, the harvesters can analyze the quality of the grain as its harvested, and make adjustments to prevent damage, providing farmers with a consistent grain quality.
The company's farmer customers tend to adapt quickly to any new software, Sanchez said.
"Farmers have actually always been to a large extent early adopters of technology," he said.
Farming is challenging, he added, and farmers move quickly to use technology that will make their lives easier or help cut costs.
John Deere tries to ensure its technology is "walk up easy," as Sanchez called it.
"For the average farmer, we want the majority of functions for a vehicle or technology easily learnable if not in minutes then in hours," he said.
In 2017, the agricultural company opened a technology office, John Deere Labs, in San Francisco, focused on furthering its AI journey and developing machine learning and AI-driven technologies. The office has been operating since then.
The AI Summit was held Sept. 25 to 26 at the Palace of Fine Arts.