[Expert Column] The Completion of Agricultural AX Lies Not in AI Judgment but in Field Execution…Connecting Agricultural Machinery and Data Is Key

[Expert Column] Lee Kwang-wook, Head of Daedong Agricultural AX Promotion TF Division

While promoting the agricultural AX(AI transformation) business, the question most often heard in the field is, “So how does farmers’ work actually change?” Even if artificial intelligence analyzes soil and crop growth data and predicts pests, diseases, and yields, if farmers must interpret the results again, make work plans, and mobilize agricultural machinery and labor, changes in the field are inevitably limited.

What farmers want is not the analysis result itself. It is to carry out the necessary work at the right time, reduce labor and production costs, and achieve a stable harvest. The competitiveness of agricultural AX is also determined not merely by the accuracy of artificial intelligence, but by how reliably that judgment is connected to actual farm work.
 

Global agriculture is moving from ‘analysis’ to ‘execution’

The movements of global agricultural machinery companies also show this change. John Deere is building a structure in which work plans established in its ‘Operations Center’ are wirelessly transmitted to agricultural machinery, and work data generated in the field is brought back to the platform. ‘See & Spray’ distinguishes crops from weeds using cameras and artificial intelligence, then selectively sprays herbicide by activating only the nozzles at necessary points. It is a method of immediately converting analysis results into equipment operations.

Kubota is also collecting data gathered from agricultural machinery in a farm management system, visualizing work status, and using it for fertilization and water management. Its goal is to connect this with autonomous and unmanned agricultural machinery to implement precision agriculture in which planning and execution, result analysis, and improvement are repeated. Kubota describes this as data-driven ‘PDCA-type agriculture’. 

These cases show that global competition in agricultural technology is moving beyond automation of individual agricultural machines toward connecting data, artificial intelligence, and agricultural machinery into a single operating system.

The remaining disconnect between analysis and work

Until now, smart agriculture has devoted much effort to deploying sensors and equipment and collecting data. As a result, the amount of information farmers can check has greatly increased, but how to convert analysis results into specific tasks and how to allocate the necessary equipment and labor remain challenges to be solved.

Soil and crop growth data are stored in different systems, and agricultural machinery and implements operate in different ways. Even when a precision agriculture solution presents a good prescription, there are cases where machinery cannot perform the work according to that prescription. Agricultural work outsourcing centers also often receive applications by phone and adjust schedules manually, so work processes and results are not left as systematic data.

To reduce this disconnect, Daedong is expanding the connection between tractors and implements by introducing tractors and smart implements that apply ISOBUS, the international communication standard for agricultural machinery. By using ISOBUS, tractors and implements can exchange work commands and data through one communication system. Prescription maps can be delivered to tractors and implements to adjust seeding rates and fertilizer application rates by location, and data such as work location, area, and input amounts can be collected again.

When the collected work data is combined with soil and crop growth data, the accuracy of prescriptions and the efficiency of farm work can be verified. The more the input amounts and results by soil and growth conditions are repeatedly analyzed, the more sophisticated the prescriptions for the next farm work become. ISOBUS tractors and smart implements are the connecting link that executes precision agriculture prescriptions and sends the results back as data.

No matter how excellent the performance of individual technologies is, if this disconnect is not resolved, it is difficult for farmers to feel the effects. The next task for agricultural AI is not to show more information, but to connect secured data to farm work and improve the next tasks based on the results.
 

An operating system that moves the field is needed

Agricultural AX is different from simply digitizing existing farm work. Data is collected from agricultural machinery, robots, drones, and facilities; artificial intelligence makes decisions needed for sowing and fertilization, pest control, and harvesting; and AI agricultural machinery, smart implements, and agricultural robots execute them. The work results are accumulated again as data and improve the next judgments.

This is also why Daedong seeks to connect AI tractors, precision agriculture solutions, drones, smart implements, and agricultural robots into a single farm work process. However, agricultural AX is not completed merely by connecting equipment and solutions. An operating entity is needed to receive requests from farms, establish work plans suited to field parcels and weather conditions, and allocate equipment and labor.

If existing agricultural work outsourcing was a method of performing tillage, sowing, pest control, and harvesting on behalf of farms according to their requests, in the future it must develop into an agricultural production operation service that manages everything from work planning to dispatch, control, and results based on data. It is a structure that connects farm applications, work plans, equipment and labor dispatch, progress status, and result reports into one platform.

If regional agricultural cooperatives, farming corporations, and agricultural work outsourcing centers use precision agriculture solutions, advanced agricultural machinery, and operating platforms, this transition is possible. This is also a way to increase access to technology for small and medium-sized farms that find it difficult to directly purchase expensive AI agricultural machinery and agricultural robots. This is because regional execution organizations can jointly operate equipment and solutions, and farms can use the work they need as a service.

The value of agricultural AX must be proven in the field

A national agricultural AX platform must also go beyond gathering data in one place and become a foundation that connects artificial intelligence judgments with actual farm work. The government must establish standards and a demonstration basis so that data, agricultural machinery, and platforms can interoperate. Companies must build integrated solutions in which hardware and software work together in the farm work process, and regional execution organizations must convert these into services that farmers can use.

The purpose of agricultural AX is not to produce more information. It is to enable farmers to carry out the necessary work at the right time and farm stably while reducing production costs and labor burdens.

Technology is already evolving rapidly. What is now needed is for institutions to open the way so that developed technologies can be sufficiently used and verified in agricultural fields and then developed further. When a structure takes root in which artificial intelligence judgments lead to the work of agricultural machinery and robots, and the results improve the next judgments, agricultural AX will finally become change in the field.

The competitiveness of advanced agriculture is not determined solely by who developed the technology first. It depends on how quickly that technology is established in the field and expanded into an industrial ecosystem. For South Korea’s agricultural AX to secure competitiveness in the global market, the government and industry must together prepare the institutional foundation for field diffusion. Now is exactly the time.