John Deere invited me to Ames, Iowa, for the pre-briefing of “JD,” the company’s new AI assistant, announced September 1, 2026. I jumped at the chance because ever since attending the opening of the Austin John Deere Innovation Hub in 2023, I’ve wondered when its investments in AI would evolve beyond embedding physical intelligence in machinery to interact directly with customers. JD is the answer. Here’s what it is, how it works, and why other companies with physical assets should pay attention. (Deere paid travel costs.)
JD AI Assistant
John Deere Operations Center is the cloud service holding each customer’s field, machine, and operational records. In operation for over a decade, the service has a wealth of data from every machine and every phase of production, e.g. tillage, planting, spraying, and harvesting. JD resides inside Operations Center and answers plain-language questions using that history along with general context about Deere products and agriculture. When is the optimal time to harvest based on historical trends? How did singulation compare across my fields, and what did it do to yield? (Singulation is how evenly a planter drops individual seeds.) Which sprayer operator covered the most acres per hour? Answering questions like these used to require digging through years of reports and raw data. “JD changes the experience from navigating through a sea of data to simply asking it a question,” said Jahmy Hindman, Deere’s CTO. “It puts advanced data analysis within reach.”
Farmer Data Commitment
The company announced the Farmer Data Commitment in parallel with JD. Ten principles cover data control, transparency, and choice. The farmer controls the data. Deere does not sell it. The farmer chooses which third parties receive it and can stop that flow at any time. Deere does not use farm data for commodity trading or speculation. “Our view is simple: farmers should control their data, clearly understand how it is used, and benefit from the insights it can create,” said Deanna Kovar, Deere’s President of Worldwide Ag and Turf.
Early Access
JD is rolling out as an early access program, with broader availability later this year through Operations Center on web and mobile platforms, with an in-cab interface in the works, as are other versions of JD for turf, construction, roadbuilding, and forestry customers.
Smarter Equipment Collects More Data
Meanwhile, the machines feeding Operations Center context are also getting smarter. See & Spray™ sprayers use arrays of NVIDIA-based vision systems in ruggedized air-cooled enclosures. The photo below shows the finned aluminum housings mounted on the back of a sprayer. Every spraying operation adds to each field’s history. As Kovar put it, every field becomes a test plot because data is collected all the time.
My take: Deere built the JD AI Assistant within Operations Center, a cloud service that customers already use, trust, and continuously feed with data. Identity, authentication, access control, and APIs are already in place, providing a rich hosting environment for AI. Although Deere did not announce agentic capability, closing the loop from answer to action is a natural extension, and I would bet the company is working on that.
This AI deployment model is an excellent fit for Deere and other companies with physical operations and cloud data for 5 reasons:
- Customer first: Deere started with the customer, not the technology. Deere designed JD to solve real-world farm problems from the start, using existing data and familiar user experiences.
- Use data as-is: No MCP servers to build, no model training, no new data lakes, no new security model.
- Trustworthy: The “Farmer Data Commitment” spells it out: “Your farm. Your data. Your call.”
- No deployment barriers: No AI tech to learn, no complicated and unfamiliar AI infrastructure to manage. Just start asking questions. And no direct cost — JD and the core of Operations Center are free for farmers. (The company does charge for PRO services aimed at equipment and fleet maintenance.)
- ROI from the start: JD delivers value from advanced analytics that farmers otherwise would not have time to do. JD makes precision agriculture ROI measurable in real-time.
In digital AI, the model comes first, then enterprises build deep stacks of customer knowledge, agents, governance, trust, and data interfaces to support it. ROI comes later, if it comes at all. Physical AI reverses the order. Start with high-value customer problems, use the data as-is, make trust explicit, and remove every barrier to using it. Using data as-is is often more practical than it sounds, particularly where a system like John Deere Operations Center already holds rich operational context and exposes it via dashboards and reports. In many cases, the same interfaces can feed the assistant.
