John Deere plans to roll out an artificial-intelligence agent in Latin America by the end of 2026, giving farmers a new way to query data from their own operations as the machinery maker expands its services business during a difficult year for equipment sales.
The tool, called JD, will allow customers to ask questions about their operations and receive answers in real time, reducing the need to navigate multiple screens, filters and reports. It will draw on data from the John Deere Operations Center, the company’s platform for monitoring and managing connected farm machinery.
The launch comes as farm-equipment manufacturers contend with squeezed farmer margins across Latin America. Deere expects industrywide agricultural machinery sales in the region to fall 15% to 20% this year from 2025, according to its third-quarter results.
In weaker machinery cycles, farmers often opt to upgrade existing equipment or software systems rather than buy new machines, Rodrigo Bonato, John Deere’s vice president of sales and marketing for Latin America, said earlier this year.
Latin America has also been an early adopter of AI in agriculture. About 26% of farmers in the region already use AI tools, compared with 17% globally, according to a McKinsey study.
JD will work similarly to AI agents such as ChatGPT or Claude, but its answers will be grounded in data collected through Deere’s own platform.
In Brazil alone, about 100,000 John Deere machines are connected to the system, according to Cristiano Correia, the company’s global vice president for sugarcane and vice president of production systems for Latin America.
Globally, Deere has more than 1.2 million connected machines operating across 520 million acres (210 million hectares).
Technology Push
The AI agent is the latest step in a technology strategy Deere has pursued for nearly a decade. In 2017, the company acquired startup Blue River Technology for $305 million, adding machine-learning and robotics capabilities that are now used across its technology platform.
Deere’s systems can already assess metrics ranging from how long a machine takes to complete a task to soil conditions for planting or harvesting. Large language models, or LLMs, are intended to make that information easier for farmers to access and interpret.
“We think about this as the language of farming”, Chief Technology Officer Jahmy Hindman said at Deere’s investor day in December.
The company has developed AI models for specific tasks, including identifying obstacles, distinguishing among tree types and detecting weeds. That technology also helps operators troubleshoot and repair increasingly sophisticated equipment, according to Cory Reed, a John Deere customer success executive.
Deere can also reuse and adapt models developed for different applications, potentially speeding the rollout of new technology.
“We estimate that each new application can reach the market in half the time of the previous one,” Hindman said.
This story was translated from the original Portuguese with the assistance of artificial intelligence and reviewed by The AgriBiz editorial staff.
