Businesses have always used the word “adoption” when talking about technology, especially when referring to what I call “freeform” technology. By this I mean tools that each user uses in their own way and at their own discretion, because their use, unlike, for example, business tools, is not subject to any workflow or explicit rules requiring them to be used in a specific way in specific cases, and can even be avoided if the user does not want to use them or does not see the point of doing so. The term speaks for itself: since we cannot or do not want to impose it, it must come from the user themselves. You can visit the archives of this blog to see how I criticized the use of this term at the time of business social networks and digital transformation (Technology adoption kills your digital transformation or Social in the workplace : forget adoption or Does driving adoption mean being off the point ?, to name but a few), and I can only conclude that we are repeating the same story, with a few subtle differences.
They talk about adoption as if it were a predictable progression, a step that would occur almost naturally between the provision of a tool and its integration into an activity. This vocabulary has long made it possible to avoid focusing on the work itself. We were convinced that equipment, training, and a minimum of support would be enough to change practices, but when we look at what is happening in the field, we see a succession of scattered trials, pilots, POCs, and a sense of agitation that leads to nothing structured. This disconnect is not new, as I showed above.
The case of AI does not bring anything new to this subject, as it merely reactivates an old mechanism that leads to confusing the use of a tool with the evolution of an activity. We have already seen this with employee collaboration tools and software suites designed to support daily work: a lot of enthusiasm at the beginning, some personal progress, and very little collective transformation. AI is just an opportunity to reiterate a long-standing observation, in the hope that we will finally draw the bottom line from it instead of believing that, miraculously, the same choices will lead to different results.
In short:
- The term “adoption” applied to technology is misleading because it suggests a natural progression towards use, without questioning the organization of work or the necessary transformations.
- The introduction of a technology, such as AI today or collaborative tools yesterday, often leads to occasional individual use without any real collective transformation or overall benefit.
- Using a tool does not mean appropriating it: the latter implies a lasting change in the way we work, both individually and collectively.
- Productive appropriation depends on organizational, cognitive, and cultural conditions that allow for a rethinking of workflows and experimentation with new team practices.
- As long as the work itself does not change, the use of a technology has no significant effect on the business, making the concept of adoption insufficient as an objective.
Adoption does not exist
The word adoption has never described what happens in a business, but suggests the idea that change can be triggered by a series of organized steps aimed at getting people to use a tool, when in fact this shift depends on how work is designed, understood, and organized. You can distribute a tool, explain how it works, and support employees as they take their first steps. None of this tells us when a team will actually change the way it works.
And this was already true before AI. All previous technologies have proven this. What is surprising today is not the existence of a gap, but the fact that we are repeating the same sequence with the same illusions. We still talk about adoption, even though this word has never helped us understand why an activity remains the same despite the novelty. AI only highlights this inconsistency.
We are already seeing the first symptoms. Users who limit themselves to using AI individually for existing tasks and old workflows without any real transformation, the explosion of shadow AI for those who are more advanced but need to think outside the box if they want to be more ambitious in their uses, and global use cases where emotional support and assistance in everyday life take precedence over professional and technical issues for the average user (How People Are Really Using Gen AI in 2025) and, ultimately, the observation of declining maturity in the face of AI (AI maturity: true progress lies in recognizing that one is not ready), which I would rather translate as an admission that it has been overrated in the past.
Using does not mean appropriating
You can use a tool without changing the way you work, and I like to draw a parallel with the arrival of electricity, which had virtually no impact on productivity until we understood that it allowed us to organize factories differently and reconfigure their layout and physical production flows.
You can achieve fast results in terms of adoption without it having any effect on a given end-to-end activity, which is the only thing that matters to the business. Progress can be made in isolation while everything else remains static and, worse still, one activity or person can single-handedly ruin all the progress made by the group (Local optimum vs. global optimum and the theory of constraints: why your productivity gains sometimes serve no purpose). This has always been true, and AI does not change this phenomenon. It just reminds us of it because it can produce instant effects that can be misleading. We believe we are making progress because we see people using AI, because some people check in substantial gains on a given task, provided that we measure it properly (How can we measure the productivity gains of AI?), but in the end, the collective and the business derive no benefit from it.
Adoption means using a tool, while appropriation means understanding its potential and even its requirements in terms of transformation, first for oneself and then for others.
Individual progress is not enough and has no significance if the interdependencies between those who are progressing and others are not taken into account to ensure that we are not playing a zero-sum game. In a team, nothing changes as long as the workflow remains the same and we are content to add up individual performances, believing that they will create collective performance. Productive appropriation, on the other hand, can be recognized when its use changes the way an individual or collective activity is performed and a new stability emerges in the way things are done.
The conditions for productive appropriation
With AI, we are rediscovering that productive appropriation depends above all on well-known conditions. We need to understand how work is organized, why a sequence unfolds in a certain way, where the dependencies lie, and what we want to achieve. Without this, no technology can bring lasting collective benefits.
These conditions are also cognitive and cultural. An organization that encourages exploration gives a team the opportunity to test, adjust, and learn, while an organization that demands perfect execution from day one prevents any evolution. Productive appropriation relies on the ability to interpret what a tool offers, to reformulate a need, and to try new approaches.
Finally, they are collective. A team must be able to review its tasks, reorganize its dependencies, and adjust its workflow. As long as everyone makes progress alone, there will be no significant improvement in activity check-in, and it is no coincidence that Moderna, whose forward-thinking vision has been much talked about this year, continues to refer to the concept of “work as a flow” (HR and IT merger: Moderna redesigns its organization for and with AI and Thinking of work as a flow: appealing, but is it realistic?). Local progress has no impact if the structure of work and the management of activities remain unchanged. Productive ownership is not the sum of individual successes, but the construction of a shared practice.
When work doesn’t change, nothing changes
When these conditions are not met, history simply repeats itself. Technology spreads and is used, but it doesn’t transform anything and therefore brings little benefit. Teams multiply their attempts, but they fail to reap the rewards of their experiments. Individuals save time, but this time is not reflected in any measurable gains for the business. There are encouraging signs, but without being able to put the pieces of the puzzle together. Collaborative tools have already experienced this situation, and Solow can rest assured: he will not become obsolete anytime soon (You can see the computer age everywhere but in the productivity statistics).
Adoption is still cited as a condition for success, when in fact it only serves to avoid talking about the problem.
Bottom Line
AI does not require us to abandon adoption efforts because it is unique, but because our experience with previous technologies should have already led us to this bottom line and made us understand that it is only a step, a necessary prerequisite, but in no way a goal or a result with which we can be satisfied. And as long as we confuse the use of a technology with the evolution of an activity, we will continue to encounter the same problems.
The adoption of a technology by a business and its employees has absolutely no value other than to create the conditions for moving on to the next stages, which are individual and then collective appropriation, and thus productive appropriation.
To answer your questions…
Adoption only ensures that the tool is used, without guaranteeing that it actually changes the activity. The article shows that transformation depends above all on how work is organized, sequenced, and understood. Training, equipping, or supporting employees does not automatically lead to a change in practices. As long as workflows, dependencies, and objectives remain the same, usage will remain superficial and without collective impact. For a business, the challenge is therefore not adoption, but the reconfiguration of work.
Like collaboration tools, AI generates initial enthusiasm and individual gains without any real collective transformation. Organizations still confuse usage with evolution of work, creating the illusion of natural progress. AI offers fast effects that mask the absence of structural change. Uses remain scattered, often individual, and shadow AI appears as soon as the tools fail to take things further. The challenge is to move beyond experimentation and rethink the activity itself.
Individual progress remains isolated until the overall workflow is reorganized. An optimized task does not improve anything if dependencies remain fixed and other steps slow down the whole process. A single misaligned link can cancel out gains made elsewhere. Without collective consistency, we end up with a sum of local performances that has no effect on overall performance. Businesses must therefore focus on structuring work rather than on individual practices.
Productive appropriation requires understanding how work is organized, identifying dependencies, and clarifying objectives. It also relies on a culture of exploration that allows for trial, adjustment, and learning. Finally, it requires collective coordination: reviewing tasks, reorganizing workflows, and stabilizing a new way of doing things. Only under these conditions can the use of AI transform the execution of activities rather than remain cosmetic. For a team, this means working together on processes.
When work remains unchanged, technology spreads without producing any transformation. Experiments multiply, users progress individually, but no measurable benefits appear at the activity level. The article points out that this situation was already visible with employee collaboration tools and persists with AI. There are encouraging signs, but they cannot be translated into concrete gains. In such cases, invoking adoption mainly serves to avoid addressing the real cause: the lack of change in the work itself.
Image credit: Image generated by artificial intelligence via ChatGPT (OpenAI)
Related articles:
Why the widespread adoption of AI by consumers says nothing about its future in the workplace
Adoption of AI in the workplace : current situation
Is AI the new Lean?
Generative AI: your organization is worth more than your tools
Nobody wants to prompt
Who is handling your artificial intelligence projects? Probably not the right people.
