An organization does not automatically become what its leaders want it to be, but rather what its tools allow, suggest, or facilitate. Indeed, when a business adopts a technology without a structured plan, it simultaneously imports a work logic designed elsewhere, by others (How management let systems do the thinking for them). Technical choices become organizational choices that no one has formulated, and while we think we are modernizing an environment, we are profoundly changing the way we work and operate, often without visibility into the possible side effects. The question is no longer whether technology influences the organization, but to what extent the absence of intention allows this influence to replace strategy (Taking back control of enterprise design: intention before tools).
This phenomenon has historically been mitigated by informal practices: employees fill in the gaps in a tool, compensate for its rigidities that have become organizational constraints, and play with its limitations (Work about work: when the reality of work consists of making things that don’t work work). As long as the pace remains moderate, let’s say “human”, the business can absorb inconsistencies, mitigate them, or work around them by using the resilience of its employees as makeshift shock absorbers. This is what happens without managers and executives being aware of it, or because they see it as a necessary and inevitable evil. But this ability to adjust erodes as soon as the speed of technology exceeds that of humans, and this is precisely what AI introduces: an unprecedented speed that makes visible and operational choices that have not been formulated for years but are very much present in the reality of work.
In short:
- Technology, particularly AI, profoundly influences organizations when adopted without clear intent, imposing implicit work logics designed elsewhere.
- AI acts as a revealer of the implicit architecture of businesses, amplifying existing inconsistencies and exposing unarticulated organizational choices.
- The absence of intentional work design allows technology to fill in the gaps according to its own logic, which can alter the operational identity of the business.
- The case of Klarna illustrates the consequences of effective technology deployment that is not aligned with the business’s identity, leading to a degraded customer experience despite correct execution.
- Organizations must clearly define their operations and operational identity before integrating AI, or risk seeing their uniqueness replaced by that induced by the tools.
AI does not disrupt the organization: it reveals its implicit architecture
It would be a mistake to consider AI as an external element that disrupts a stable system. In reality, it plays a revealing role: it absorbs what it observes, almost mimicking it, in order to replicate it. This phenomenon occurs without any intention on the part of the technology; it is simply an amplification mechanism that causes organizational ambiguity to be transformed into an applied and repeated directive, or a poorly designed process to have its effects accelerated and amplified.
This process highlights a crucial fact: a business that has never intentionally defined how it operates allows its technical artifacts to do so for it, and it is important to realize that this is not a future risk but a mechanism that is already in motion. Every omission in the design of work and the business becomes an opportunity for technology to impose its philosophy, its assumptions, and its vision of the organization and work. The resulting dynamic is implacably logical: what is not decided by humans will be decided by the system that executes.
Technology that accelerates is technology that exposes
For a long time, organizations have sought to optimize their operations without questioning their foundations. They adjust a process, improve a workflow, or reorganize a team, but they do not revisit the principles that should structure the whole. AI accelerates operations, but above all, it accelerates the consequences of this lack of clarification. The machine applies without nuance what the business has left unresolved, but that’s not all. A poorly designed process that worked thanks to employee compensation thus becomes a time bomb if it is entrusted to AI without being re-examined. Gray areas are interpreted and become rules that no human has validated, inconsistencies are systematized and accelerated, and any design errors are set in stone more than ever before.
AI, like all technology before it, will not create new problems, but by accelerating things on a large scale, it makes them more visible than ever and their consequences, when harmful, now unbearable (The new process excellence: Taking AI from toy to transformation).
This exposure makes it impossible to hide unmade choices or imperfections that have been allowed to persist. The inertia that once mitigated the imperfection of systems no longer works because AI acts faster than the human capacity to rebalance. But it is not the technology that is malfunctioning, because what it produces is less the problem than the symptom of the problem, which is that the business is confronted with the reality of its own design, often much less perfect than it thought.
Klarna: everything worked as planned, and that’s precisely the problem
The Klarna episode has been widely discussed (When AI Turns Your Secret Sauce Into Ketchup), but it should be taken for what it is: the logical result of a design that was not fully implemented. AI didn’t invent anything; it simply did exactly and perfectly what it was designed to do, and everything worked perfectly. The problem was neither in the technology nor in the execution, but in the design, where the “identity” dimension had been completely overlooked (Efficiency vs. uniqueness: the false dilemma of operations).
So everything worked perfectly, except for one detail: what customers were experiencing may have been an experience that worked, but it wasn’t the Klarna experience, not the one they were used to, not the one they were promised.
Here, AI took (some) control of customer service operations and, above all, their tone, what they conveyed or rather did not convey, because AI, like any technology, has no intention and no personality unless this is planned in advance. AI did not fill the void left by the absence of design, and customers left.
Klarna understood perfectly what they wanted to achieve in terms of efficiency and quality, but they simply forgot who they were when it came to implementing the technology, instead of starting by (re)designing their operations with this factor in mind.
Operational identity as a long-neglected strategic variable
Most businesses talk a lot about identity but very little about operational identity, even though it is this that determines how an organization distinguishes itself in its daily practices. It is not a slogan or a stated value, but the way in which work is carried out. Two businesses in the same sector may sell the same thing, but their operating styles sometimes differ radically, and it is this difference that creates the uniqueness perceived by both customers and employees.
Standardized technologies have already begun to reduce this uniqueness because they impose models of action that do not take internal specificities into account. AI can amplify this process if the business does not establish a rigorous framework to formalize what it wants to preserve. The challenge is not to contain the technology, but to define the internal framework that will guide its use, otherwise it will impose its own logic.
Designed for AI or designed by AI: a decisive choice
We can summarize things as follows.
An organization “designed for AI” first defines itself. It spells out its principles of action and the benchmarks that structure its way of working, and clarifies what makes it unique in its execution. Once this foundation is in place, it incorporates technology so that AI aligns with an already established identity rather than suggesting a substitute. In this configuration, technology “augments” a chosen framework rather than creating one by default.
An organization “designed by AI”, on the other hand, adopts technology before formalizing what it expects of itself. Technology then fills in the gaps. Unmade decisions are left to its discretion, and what was once part of the operational identity is interpreted or even invented by the machine based on the signals it observes or its standard configuration. Part of the design therefore becomes an emanation of the tool and its philosophy, which ends up occupying the place that a formalized identity should have held.
This distinction determines the path a business will take, depending on whether it puts technology at its service or allows it to take control of the way it operates, and therefore of the way customers and employees experience it, and ultimately of the brand and its perception.
Before a way of working takes hold that owes nothing to the business’s decisions
The business now has a narrow window of opportunity to regain control. It is not a question of mastering the technology, but of mastering its own organizational design. What the business does not clarify will be decided elsewhere, and perhaps with no turning back, given the seemingly significant potential for acceleration and transformation offered by AI. The urgency lies not in the deployment and adoption of technology, but in clarifying what the business wants to assert in its way of working so that AI fits into a framework that resembles it (Enterprise design before architecture: putting the company back the right way up).
Now is the time for business to define what they want to be, how they want to do things, and why, otherwise it will be imposed on them.
Bottom Line
The question today is not whether AI will transform organizations, because its introduction is already leading to organizational adjustments, some subtle, some more visible, even without explicit intent. The challenge is to determine whether the form of the organization is the result of a conscious choice or whether it emerges from technology when no intention has been stated.
As long as the intention remains unclear, technology takes over as the arbiter, and the challenge is therefore to regain the initiative before day-to-day operations are shaped by a dynamic that no one has formalized, sometimes contrary to expectations.
AI does not (yet?) seek to take the lead, but nature abhors a vacuum, so it imposes its mark where the business has not defined its own. To retain its uniqueness, an organization must now clearly set out the principles that guide its actions, the areas where it wishes to delegate, those where it refuses to do so, and how it articulates the whole. It is in this ability to write its own framework that lies the consistency between what it claims to be and what it becomes.
More than a problem, this is a unique opportunity for businesses to (re)ask themselves the right questions before a major shift.
To answer your questions…
When a business adopts a technology without clarifying what it wants to preserve or change, it allows the tool to structure its operations. Employees often compensate for the limitations of systems, but AI, which is faster, quickly exposes gray areas and inconsistencies. What has not been decided then becomes an implicit rule that is applied mechanically. To maintain control, the business must define its principles of action before integrating a technology, so that the technology amplifies a clear intention rather than replacing it.
AI strictly applies what it observes within the organization. It replicates practices, including shortcomings or ambiguities that humans had previously compensated for. By speeding up execution, it reveals structural flaws that have become impossible to hide. The problem then stems less from the technology than from the lack of explicit organizational design. It is essential to reexamine processes before delegating tasks to AI to avoid solidifying existing errors.
At Klarna, AI performed perfectly for what it was configured to do, but without incorporating the tone and uniqueness of the brand experience. This gap in operational design was filled by the tool’s standard logic, creating an experience that was functional but not true to Klarna’s identity. This shows that technology can be technically successful but strategically unsuccessful if the business has not defined what it wants to preserve in its customer relationship.
Operational identity reflects the actual way of working, beyond slogans. It creates the uniqueness perceived by customers and employees. Standard technologies tend to standardize practices, and AI can accelerate this homogenization if nothing is defined internally. By clarifying what makes it unique in its execution, a business can integrate AI without losing what sets it apart. This is a strategic challenge in maintaining consistency and differentiation.
An organization “designed for AI” first defines its principles, identity, and operating methods, then integrates AI to reinforce this framework. Conversely, an organization “designed by AI” adopts the technology before clarifying its choices, allowing the tool to fill in the gaps and guide its practices. This distinction determines how internal change is managed: either the business guides the technology, or it is subject to its default logic.
Image credit: Image generated by artificial intelligence via ChatGPT (OpenAI)
