Mr. John Deere started a revolution with a simple, but impactful invention - he designed a non-stick plow 190 years ago (1837), a significant agricultural innovation that helped farmer productivity. Since then, the company has continued to innovate for ~ 2 centuries in the agro-equipment and construction space. A big piece of this is about using data (farmers have a lot of it), precision agriculture, autonomy and physical AI solutions to create complex agricultural machines from seeding and irrigation, to pest control and harvesting.

Deere has a 35 year old history in collaborating with Iowa State University to integrate academic research into practice. It recently organized a Tech Day at the Iowa State campus in Ames, to discuss how this collaboration is impacting the business of farming through use of agricultural science, precision AG, AI and data analysis. A key announcement was the launch of JD.Com , an artificial intelligence (AI) agent to analyze and learn from historical field and equipment data. Based on site & crop specific data provided by individual farmers for their geography and crop type, this agent can answer questions like historical fuel or fertilizer usage, equipment performance, and advice for the upcoming planting season.

John Deere and Iowa State - A Unique Collaboration

The Tech Day was conducted on September 1 at the Ag Innovation Laboratory in Ames, Iowa. Iowa State’s stated purpose is “Science with Practice”, which demands deep collaboration with industries like John Deere. The partnership is unique in the way intellectual property (IP) is developed and used. In exchange for up-front funding, Deere has an exclusive license to use this IP for agricultural products. This model is unique and serves as an example of how industry-academia collaborations can be structured in other areas of medicine and engineering.

Deere CTO Jahmy Hindman and Matt Darr, Director of the Ag Innovation Laboratory and an Endowed Professor at Iowa State, led off the festivities with a presentation that covered four major areas of recent collaboration:

  1. BoomTrac™: keeps the boom that houses sprayers on a gantry perfectly positioned in terms of spray area and height. ExactApply™ Multi-Rate delivers multiple rates across all 11 boom sections in one pass – placing the optimal spray rate in the right zone with precise pulsing to match nutrients and pesticides to field conditions, and maximize the impact of every input. Field data maps (conditions, weeds, plants, historical yields) are analyzed to provide variable spray rates, with savings of up to 80% in materials and reduced pesticide application in high yield areas. Other technology blocks like AutoPath™ (Rows) and AutoTrac™ Vision 2.0 plan the sprayer path and keep it centered on the row at higher speeds, reducing row damage and lost yield (Figure 1). A recent report indicates that almost 50% of farmers are sickened by pesticides every single year. Using precision techniques increases worker safety, and creates sustainability and environmental benefits.
  2. Harvest Settings and Predictive Ground Speed Automation: Farmers can use a dashboard application to specify product quality variables like grain loss, foreign material content and amount of broken grain during harvest. The harvesting combine internally adjusts machine parameters like fan and motor speed, and threshing, sieve and chaffer clearances to optimize user-defined parameters. Predictive Ground Speed Automation uses satellite imagery and real-time camera data to predict crop and weed biomass and variability ahead of the combine, enabling optimal speeds and yields.
  3. ExactEmerge™: seeding is a critical piece of the farming journey - seeds need to be planted within a two week window, at the optimal depth, correct spacing and appropriate surrounding soil texture to enable maximum harvest yields. Deere’s ExactEmerge ™ combiner (Figure 2, Top) incorporates up to 36 planters that traverse the field, create precision furrows, plant seeds and dispense the right amounts of fertilizer, pesticide and water at each location. Older technology mechanical planters moved at 3 mph, constrained by the dynamics of the vehicle as well as the planting speed. ExactEmerge ™ incorporates an electric drive - the combination of this and FurrowVision ™ triples the speed to 10 mph, enhancing productivity during the short planting window. Other benefits include elimination of excess resources like fertilizers, pesticides and water, with associated cost and sustainability benefits.
  4. FurrowVision™: is a sensor mounted at the wheel base (Figure 2 Bottom) which constantly measures the depth of the furrow using a combination of cameras and a green laser imager. It monitors seed placement and furrow conditions in near-real-time and displays it via a live feed stream on a display in the cab. The operator can then make adjustments to the downforce applied. It is critical to control the furrow depth through the downforce since higher or lower levels can reduce yields dramatically. The target depth relies on learnings from years of data and experience.

AI and Data - Helping Farmers Optimize Operations

At its heart, John Deere is an equipment company that has an intimate understanding of what farmers need and how to support those needs. Complex electro-mechanical and materials expertise is integrated to produce the tools that farmers need - to plough, seed, fertilize, protect and harvest crops - reliably, economically and sustainably. The ultimate example of this is the harvest combine, essentially an autonomous factory on wheels that harvests, threshes, cleans and separates food-grains like corn via a hopper into a loading truck for immediate dispatch to end customers. LED lighting, GPS connectivity and camera sensors enable 24x7 operation, critical when harvest cycles are ~2 weeks. Precision technologies like Harvest Settings and Predictive Ground Speed Automation can be integrated into the X9 operation. BoomTrac™, ExactEmerge™ and FurrowVision™ are integrated into pre-harvest operations like seeding and spraying.

AI tools are springing up in all domains - John Deere’s magic sauce is the ability to integrate complex physical machines with precision Ag technologies, supported with learning from years of data to truly help farmers optimize operations every year in the face of variable climate, workforce availability and market forces.

John Deere already provides an Operations Center™ technology which enables farmers to digitally store years of operational data into a structured format. It is an online farm management system that enables access to farm information anywhere and anytime, through web, tablet, or phone. Farmers can review this and make decisions on current and future operations, optimizing every pass, to reduce costs, and improve profitability and worker safety (Figure 4):

Apart from the display interface, other pieces of hardware and connectivity like the StarFire™ Receiver (tracking exact location of equipment on the field) and the JDLink™ Modem (maintains equipment connectivity to the Operations Center™) provide farmers with complete remote and in-cab operational visibility in near-real time. Currently, Operations Center™ is accessible in 100 countries to all farmers (John Deere customers or not) where regulations authorize farm equipment connectivity. Amazingly, it is a free service.

JD ™ is the next evolution of the Operations Center™. The AI agent learns from individual farmer data that has been populated in the Operations Center™. It uses a combination of LLMs (Large Language Models) for verbal data and neural networks for quantitative data to help farmers turn years of field, machine, and operational data into practical operational insights.

Data security, ownership and privacy are key considerations - individual farmer data is protected and used only for that particular farmer. A Farmer Data Commitment plan is in place, with the following essential principles:

  1. You control your farm data.
  2. John Deere does not sell your farm data.
  3. John Deere uses your farm data only as described in its agreements and policies and will communicate changes clearly before they take effect.
  4. John Deere will require dealers and connected partners to maintain transparent data practices.
  5. John Deere uses your farm data and aggregated, anonymized data to deliver measurable value to you through improved machine performance and decision-making insights.
  6. You can choose which third parties to share, and not share, your farm data with.
  7. You have the freedom to turn off the flow of your farm data to third parties at any time.
  8. John Deere does not use your farm data for agriculture commodity trading or speculation.
  9. You should receive clear value from your farm data.
  10. John Deere will continue to listen to farmers, strengthen its data practices, and develop new solutions that help you receive value from your farm data.

Initially, JD ™ will be rolled out to selected customers in selected markets. The plan is eventually to launch it in all markets which support Operations Center™ capability. The current plan is to not charge a fee (for customers and non-customers), with the intent being to make it a sticky ecosystem for customers and an attractive feature for new customer conversions. Since the AI agents have to be individually trained for each customer, the investments in cloud and compute infrastructure could be expensive, something that John Deere is prepared to make for earning goodwill, retaining existing customers and attracting new ones.

John Deere has a legacy of innovation in designing and building complex electromechanical equipment for construction, landscaping, agriculture and forestry. The company is transforming itself into a data and AI innovator for these sectors, integrating physical equipment with precision technologies, data and AI-based learning to help farmers feed, clothe and energize a growing world - sustainably, efficiently and safely.

Sabbir Rangwala is President of Patience Consulting LLC which owns the Autonomy of Things (AoT®) Trademark