Agility in Digital Farming

Editor’s note: This article is part of a series focused on digital farming. These articles represent the body of interest that Michael Collins has witnessed and collected. Please let him know what you think of this topic, or if there is a need for clarification, by leaving a comment below or e-mailing him at [email protected]. Future articles will focus on “Latency in Digital Farming” and “Security in Digital Farming”.


Digital farming is developing in a market and a world where there is rapid and unpredictable change. As with any new technology, our detailed knowledge early on in the market development is limited, as is the state of the software. We expect to learn a lot about the technology and the state of the digital twin, and each of these will involve changes that will have to be made. Whether the change is in the processing of the weather, the crop, or the market, or in the technology of the digital twin, the digital twin is emulating the cultivation and alarming for decisions that need these changes at that point in time.

In previous articles, we highlighted many of these areas like processes, decision loops, data, user interfaces, etc. In our understanding of digital farming, the rapid and unpredictable changes may be several hundreds that are generated in the digital twin on any given farm. How are these changes going to be incorporated in the farm? The farm and its cultivation is being examined, analyzed, and alarmed upon continuously at one or hundreds of individual locations, each which may be for only square meters of virtual soil, the productive assets.

The digital twin and the digital platform upon which it runs break the cultivation down into many, many virtual machines. Each virtual machine is running asynchronous to each other. In some cases processes represented by other virtual machines are running asymmetrically, based upon many different processes underway at any time. Again, the digital twin can be looking at millions of pieces of information in the asset bases, each of a size of potentially only one square meter. This digital twin expends, effectively, no additional effort to run one or a million asset processes, continuously and repetitively! And, each alarm represent a potential decision that must be actioned.


What is important is that even though each productive asset is being investigated continuously and repetitively, it is still an integral part of the whole: the field. With all of the potential changes in processes, inputs, actions, stresses, etc. each productive asset is being investigated and examined individually and collectively. As our system matures or the cultivation progresses our refinement in the understanding of the needs for the decision processes must be reintegrated continuously.

There is a new management paradigm that enables organizations and systems to manage, thrive, and even anticipate the changes in the farm and the digital twin and to scale to every productive asset of that farm.

It enables a team, a unit, or an entire enterprise to nimbly adapt and upgrade products and services to meet rapidly changing technology and customer needs with efficiency gains, quality improvement, or even completely new products and services. 

Competitive advantage flows not from the technology itself but rather from the agility with which organizations understand and adapt the technology to meet customers’ real needs. 

Both quotes are from the forward of The Age of Agile by Stephen Denning.

This concept, originally distilled from software development, is called “Agile”. In this regard, it will apply to our digital twin and since the digital twin emulates the cultivation(s), and vice versa, the cultivation should emulate the predictions of the digital twin. Agility in digital farming is much like the evolution of the term in software development, but it applies most directly to the productive asset of the soil and stewardship, decision processes, and data that support this asset.

Agility from Wikipedia is a human condition and by definition:

Agility or nimbleness is the ability to change the body’s position (or form) efficiently, and requires the integration of isolated movement skills using a combination of balance, coordination, speed, reflexes, strength and endurance.

However, here it is now being applied to software development, and more and more to organizations. In these instances, there is wide disparity in definition, but there are commonly recognized concepts and meaning.

In the literature, for software, agility is a state of thought that conceptually converges around small teams, small tasks, and short timeframes. The term is being used to represent, and mean, nimble, balanced, coordinated, fast, reflexive, strong, and enduring, but at the heart of its meaning it is about the form of the system, the nature of its malleability, and the speed of response.

In order to really understand agility, understanding the form of the system is critical. The form of the digital twin must align with the form of the emulated reality and the nature of the work. The form of the system demands a structure that enables the converged concepts of small teams, small tasks, and short times. This convergence will enable continuous delivery and decision making, but it will concurrently require more software modularity and integration in the architecture of the need, the material solution, and the organizational solution.

A central concept in this discussion of agility, that heretofore has not been discussed, is this concept of architecture. All systems have a form, and the form of these systems is their architecture; intended or unintended, an architecture exists. This form by extension is top down, bottoms up, and across the different aspects of the system (i.e., need, material, and organization).

What we take away from this statement is that there are some immutables. The first immutable is that there is a need. The expression of this need is the operational form of the system and this form emulates the concept of the real world and the actions to be taken. The second immutable is that there are many solutions that respond to this need, and there is a one-to-many mapping of the need to each of the potential solutions that could respond. The third immutable is that there is a material solution and there is a complementary organizational solution. For agile, this enables small tasks, small teams, and quick response in the alarming and delivery of continuous decisions.

A key principle about this mapping is that there are many complementary organizational and material components to each solution. Further, each of the need and complementary organization and material components decomposes. They are all three hierarchical and every element is linked peer to peer, in-sequence, and parent to child. This linkage maintains the required integration to the lowest identifiable elements. This means for every need there is an equivalent material component and organizational component to the solution for that need.

By extension, there is also a direct correlation between the nature of work and the digital twin that emulates it. The digital twin will identify small tasks (needs), for small teams to model or emulate, and they will do this over a short period of time. This effort at the digital twin will directly relate to the small nature of the task (a digital soil asset), with a small team to accomplish the task and a short period of time (inside the decision loop of the process).

The above summary of this synthesis of the system truly requires the nimbleness, balance, coordination, speed, flexibility, strength, and enduring power of an agile environment. Even more it requires a well-defined architecture of what is and what is not a part of the system.

Leave a Reply