88% already use AI. Why do only 6% see results?

88% of companies already use artificial intelligence. Only 6% see a real impact. The gap isn't technological: the technology works. What hasn't changed are the organizations surrounding it.

Before talking about pillars and methods, it's worth pausing on a more uncomfortable question: why, with the technology already available and working, do most companies still fail to see results? The answer is almost never the tool. It lies in fear, communication, and the absence of a real method behind adoption.

What fears actually emerge when a company incorporates AI?

When an organization announces it will “adopt AI,” the first thing that activates isn't always curiosity: often, it's insecurity. People don't hear “we're going to be more efficient”; they hear “am I going to be replaced?”. That fear is rarely spoken aloud in a meeting, but it seeps into every decision: who shares their ideas and who keeps them to themselves, who tries something new and who prefers not to take risks, who collaborates with the tool and who resists incorporating it.

There is a second fear, less talked about but just as real: the fear of being exposed. If AI makes visible how long a task actually took, or how much of that workload was accumulated inefficiency, some people feel their value is at stake. And there is a third, more structural one: the fear that the company will use technology as an excuse to demand more, with the same resources and the same time, only now “AI-assisted”.

None of these fears are resolved with an internal memo or a two-hour training session. They are resolved—or deepened—depending on how the people leading the change lead.

What mistakes do leaders make in communication?

The most common mistake is treating AI communication like a product announcement instead of a conversation about people's work. The tool is introduced, productivity metrics are shown, and it's assumed that the leadership's enthusiasm will be contagious. That doesn't necessarily happen. When the message arrives without acknowledging the real uncertainty it generates, it can be interpreted as a lack of honesty rather than confidence.

The second mistake is prolonged ambiguity. Leaders, often out of caution, avoid clearly stating what will change in roles, metrics, or expectations. That ambiguity doesn't always calm things down; it can fuel rumors. People construct their own version of the story, and often that version is worse than reality.

The third mistake is communicating only once and moving on. AI adoption is not an announcement; it's a transition. And transitions need repetition, space for uncomfortable questions, and evidence that the message doesn't change depending on who is being spoken to. When the discourse toward the leadership team is different from the one reaching operational teams, trust can break before the tool is even used.

Why isn't training alone enough?

Training usually solves part of the problem: how to use the tool. It teaches buttons, prompts, and functions. What it doesn't necessarily teach—nor can it solve in a one-day workshop—is which decisions now belong to a person and which can be delegated, what a “good job” means when part of it is performed by a machine, or what to do when the process surrounding the tool remains exactly the same as before.

Incorporating a new capability on top of an already broken process doesn't fix the process: it can accelerate dysfunction. If the working system, the way information flows, who decides what, and how performance is measured do not change alongside the tool, training can produce people who know how to operate AI but continue working in the exact same way as before.

Training builds skill. Trust and real change are built elsewhere.

How is trust built during adoption?

Trust is not declared: it is demonstrated through accumulated evidence. It is built when people see that early AI experiments were small and bounded, and that a result that didn't work was treated as learning rather than failure. Also when it becomes clear, in writing and in practice, which decisions remain human, especially those involving judgment, direction, and consequences for people.

It is also built by giving real autonomy, not just authorization. It's not enough to tell someone “you can use the tool”: you have to give them space, time, and backing to try it without an early mistake being used against them. And it is sustained with transparency about what is being measured. If speed or volume suddenly starts being tracked in a way that didn't matter before, people notice. If the reason isn't explained, they may interpret it as surveillance.

When these conditions are present, adoption stops feeling like something imposed and starts feeling like something built together.

What changes when implementation has a method?

Without a method, AI adoption can turn into a collection of loose initiatives: one team tries a tool, another buys a different license, and a third waits for instructions that never arrive clearly. Everyone moves at their own pace, without a shared logic, while fears, communication mistakes, and adoption difficulties multiply without a comprehensive view.

With a method, the order of priorities changes. First, the way of thinking about change is addressed; then, it is decided where and how to incorporate AI to generate real value; and finally, the work system—structure, authority, incentives, and processes—is redesigned to sustain that transformation over time. The method doesn't eliminate uncertainty, but it offers a structure to manage it.

Three pillars that must move together

Faced with these questions, our answer is not a tool or a course. It is an approach based on three pillars that must advance together, because none works separately.

Cultural mindset shift. Before any structural or technological change, the people living through the change need to be able to view it differently. This means stopping treating transformation as a project with an end date and starting to consider it a permanent capability. It implies understanding the organization as a living system and designing culture, leadership, and recognition considering that people continue to be the source of value. Well-managed change should empower people, not exhaust them.

Product-focused AI. AI is not simply a feature added to the existing organization. It can become a lever to operate in a context of constant change. This implies redefining what high performance means, observing work flow instead of just measuring hours, and freeing people for tasks where their judgment, creativity, and decision-making capacity generate the greatest value, while AI absorbs volume and repetition. Teams well-equipped with AI can significantly expand their capacity, forcing us to rethink how we distribute tasks, roles, and talent. It also implies incorporating innovation into daily routines instead of treating it as an exceptional program.

Work system. Every organization is a system, and every system is designed to produce the results it currently produces, whether that was the original intention or not. People, structure, authority, direction, capital, and workflow are interconnected and determine much of what a company is capable of producing. When a single piece is modified without observing the others, consequences that were not always foreseen appear. Therefore, system redesign cannot be a final detail: it must be thought out from the beginning alongside cultural change and the incorporation of AI.

Choosing a path, not a destination

The question is not whether an organization will change. It will do so, with or without a method. The real question is whether that change will empower the best in people or wear them down, and whether it will be built with intention or simply left to happen.

There has never been a better time to evolve with purpose.

--