AI: The First Step
Starting Point
In the previous article (AI: The Efficiency Trap), a central issue was raised that often goes unnoticed: when an organization approaches artificial intelligence solely from a rationalization perspective—cost reduction, automation—it may improve certain operational indicators while simultaneously weakening its real ability to generate value.
This apparent contradiction is not caused by the technology itself, but by the way decisions are made. The system is being intervened without distinguishing the role each of its parts plays, implicitly assuming that everything can be optimized under the same logic. However, that assumption is precisely what introduces risk.
For this reason, before deciding how to apply AI, a prior step is required. It is neither technical nor organizational. It is conceptual.
It is about classifying the operation.
The objective is not to make what exists more efficient, but to understand which parts of the system should continue to exist, which should be transformed, and which no longer make sense.
Classifying Before Acting
Every relevant decision regarding AI ultimately depends on the type of element being addressed. Without that distinction, even well-executed initiatives can produce counterproductive effects simply because they are applied in the wrong place. Classification is therefore not a theoretical exercise, but a necessary condition for acting with judgment.
This need extends to both the processes that enable the organization to function and the products or services it offers to the market. Both belong to the same system and are deeply interconnected. Attempting to optimize one without understanding its role within the whole leads to partial decisions, where one component improves in isolation while the overall system is affected.
Only after this initial classification can decisions begin to be made consistently.
Three Types of Elements
When the operation is observed from this perspective, a distinction emerges that fundamentally changes how intervention should occur. Not all processes and not all products serve the same function within the system, and that difference is precisely what should guide any action.
Some elements belong to the core of value. These are the ones that define the value proposition toward the customer, contain critical knowledge, or sustain the organization’s real differentiation. This is where the true capacity to compete resides. Intervening in these elements without deep understanding may improve efficiency in the short term, but at the cost of eroding what makes the organization relevant. In these cases, the objective is not reduction, but preservation, strengthening, and, when necessary, careful redesign.
Other elements fulfill a different role. They are necessary for the system to operate but do not differentiate it. Their value lies in efficiency, reliability, and cost. These are natural candidates for automation and standardization, where AI can deliver significant gains without compromising the essence of the business. Failing to act here is not neutral—it leads to a growing competitive disadvantage.
Finally, there are elements that have lost their original justification. They persist due to organizational inertia, outdated constraints, or decisions that were never revisited. Although they continue to consume resources, their contribution to value is minimal or nonexistent. In these cases, the question is not how to optimize them, but why they still exist. The appropriate action is not to improve them, but to eliminate or completely rethink them.
The Consequence of Seeing Differently
Introducing this distinction is not merely descriptive; it has direct consequences for decision-making. Once it is recognized that not all elements are equivalent, it becomes evident that not all interventions can follow the same logic.
The most common mistake arises precisely when this differentiation is lost. Organizations automate processes that belong to the core of value, weakening their competitive capacity. They leave support processes untouched that should be optimized, accumulating inefficiencies. And they invest in improving activities that no longer make sense, prolonging structures that should have disappeared.
In all these cases, the technology may be correctly applied from a technical standpoint, but incorrectly positioned from a strategic one.
The problem, therefore, is not failing to use artificial intelligence. It is using it without distinguishing where and for what purpose.
What Actually Sustains the Operation
So far, the focus has been on what is visible: processes and products. However, none of these elements exist in isolation. Each one depends on a less visible but equally critical foundation that makes the operation possible.
Understanding that foundation is essential, because any intervention in processes or products directly affects what sustains them. Ignoring this dimension leads to incomplete decisions, where the surface is modified without understanding the underlying structure.
The Assets That Generate Value
The value of an organization does not lie solely in what it does, but in the assets that enable it to do so consistently. Among these assets, information plays a central role, as it defines the ability to understand the environment, anticipate behavior, and make informed decisions. It is not merely about accumulating data, but about interpreting and using it effectively.
Relationships determine access to the market and the quality of interaction with customers, suppliers, and internal stakeholders. They are built over time and rely on trust, making them difficult to replicate. Experience represents the accumulated learning derived from real operations, the practical knowledge that allows organizations to handle situations beyond formal procedures. In some cases, it is distributed across the organization; in others, it is concentrated in a few individuals, which introduces additional risk.
Knowledge itself is not homogeneous. There is general knowledge that can be acquired or transferred relatively easily, and there is specific knowledge that emerges from context, practice, and organizational history. It is this specific knowledge that truly differentiates.
Alongside this, key individuals concentrate significant portions of that experience, relationships, or knowledge. Their role is often not fully reflected in formal structures, yet their impact is decisive. Finally, infrastructure—both physical and digital—enables the operation to take place. What matters is not only ownership, but also availability, flexibility, and efficiency of access.
How These Elements Connect
Once these assets are identified, it becomes clear that they do not carry the same weight across all types of processes. The core of value depends heavily on specific knowledge, accumulated experience, relationships, and the individuals who sustain them. Intervening in these processes therefore means intervening in those assets as well, which requires particular care.
Support processes, by contrast, rely primarily on structured information and infrastructure, which explains why they are more suitable for automation without compromising the essence of the business. Obsolete elements, on the other hand, tend to rely on inherited structures that have lost their justification, so their removal simplifies the system without affecting its real capacity.
Understanding these relationships makes it possible to anticipate the consequences of any intervention before it is carried out.
Deciding with Judgment
Only after going through this process is it possible to make decisions with clarity. Each process and each product requires an intervention that reflects its nature and role within the system. In some cases, what exists must be preserved; in others, strengthened; in many, optimized; and in others, eliminated.
The key is to abandon the idea that there is a single correct way to intervene. The appropriate action does not depend on a general rule, but on a precise understanding of each element and the assets that sustain it.
The Transition
All of this takes place while the organization continues to operate. There is no opportunity to stop the system and redesign it from scratch, so decisions must be implemented progressively, maintaining operational continuity.
In practice, this resembles changing engines in mid-flight rather than building a new system from the ground up. It requires balancing transformation with stability, ensuring that the system continues to function while it evolves.
What Comes Next
This establishes the first step: understanding what exists within the organization and what sustains it.
However, this understanding is still qualitative. The next question is how to perform this classification consistently and how to decide, in each case, what action should be taken.
Answering that question requires moving from general criteria to a structured method that allows evaluation, comparison, and prioritization.
That is the next step.