From change management to changefulness

I have been thinking about change management lately.

Not because I have suddenly developed a deep longing for stakeholder maps, communication plans, training schedules, and adoption dashboards. Although there is, admittedly, a certain comfort in a well-structured implementation roadmap.

I have been thinking about it because Deloitte’s 2026 Global Human Capital Trends report makes a point that feels increasingly difficult to ignore: the words “change management” and “training” may no longer be fit for purpose. According to the report, only 27 percent of respondents believe their organizations manage change effectively, and only 8 percent believe their organizations are highly effective at meeting the continuous, always-on learning needs of their workforce. Deloitte argues that organizations need a new vocabulary focused on growth and adaptiveness.

This points to something many of us experience in practice. Change has become continuous, distributed, and deeply embedded in the work itself to be managed primarily as a program.

This does not mean that change management is irrelevant. But it does mean that some of our familiar models now feel too slow, too linear, and too confident in their own assumptions.

ADKAR and the illusion of managed change

ADKAR is a good example. It is simple, memorable, and useful as a basic mental model. But in my view, it is not much more than that.
I have learned about ADKAR, and I have been in organizations where it was used. My experience has almost always been that real change is far messier than the model suggests. People do not move neatly from awareness to desire, then to knowledge, then to ability, and finally to reinforcement.

They move back and forth. They interpret. They resist. They experiment. They comply on the surface. They misunderstand. They create workarounds. Sometimes they invent better ways of working than the official program had imagined.

And that was before AI!

In an AI context, ADKAR becomes even more problematic. Not just insufficient, but in some ways an illusion. It suggests a level of sequence and managerial control that simply does not match the reality of AI-driven change.

AI is not one change with a beginning, middle, and end. It is a continuous disturbance. It changes how we write, search, analyze, code, decide, communicate, learn, and create. It changes what good looks like. It changes what speed feels like. It changes what expertise means. It changes what junior people can do, and what senior people are expected to contribute.

While an organization is still creating awareness around one AI use case, the tools have changed, the model has improved, the risks have shifted, and someone in the organization has already found a completely different way of working.

That is why I have already changed my mental model. And I would encourage others to do the same. Not because structure is bad. But because the most useful structure today is not a rollout plan pretending that the future state is fully known. It is a learning system that helps the organization move, sense, adjust, and improve while the future state is still emerging.

The organization is already moving

Much classical change thinking still carries the shadow of the old unfreeze-change-refreeze metaphor: Prepare the organization, make the change, and then stabilize the new state.

It is a powerful image. But I am not sure it fits the world we are now in.

Most organizations are not frozen. They are already in motion. Customers are changing expectations. Employees are experimenting with new tools. Competitors are moving. Technology is advancing. Regulation is catching up. Skills are becoming obsolete. New opportunities appear before the old roadmap has been delivered.

So the leadership task is no longer to unfreeze a stable organization, change it, and then freeze it again. The leadership task is to create enough stability inside an already moving organization for meaningful change to happen.

That is a very different challenge.

It means we should stop treating status quo as the starting point. In many organizations, status quo barely exists. What exists is a stream of ongoing adaptation, local improvisation, hidden workarounds, small experiments, and competing interpretations of what matters.

In that reality, meaningful change is not created by pretending we can pause the organization while we transform it. We cannot. But we can create direction. We can decide what must remain stable so everything else can move. We can build shared language, better conversations, faster feedback loops, and a stronger sense of what good looks like.

From adoption to adaptive capacity

This is why Deloitte’s idea of “changefulness” is so useful.

The report describes changefulness as the ability to adapt, experiment, learn, and evolve as a daily muscle embedded in work, not as a disruption. It also finds that organizations that successfully cultivate this adaptive approach are 2.4 times more likely to report better financial results and provide more meaningful work.

Changefulness is not just more change. In fact, it may be the opposite. It is a way of reducing the drama of change by making adaptation part of the operating model. Not as an extra layer. Not as another transformation program. But as part of how the organization creates value.

That shift matters deeply in the age of AI. If we treat AI adoption as a traditional change program, we risk focusing too much on communication, training, and adoption metrics. How many people have completed the course? How many have logged into the tool? How many use cases have been registered?

Those things may matter. But they do not tell us whether the organization is becoming more capable. They do not tell us whether people are making better decisions. Whether teams are redesigning work. Whether managers understand the new risks. Whether AI is strengthening judgment or bypassing it. Whether people are becoming more adaptive or merely more productive on the surface.

Awareness is not enough when people are already experimenting before the official program begins. Desire is not enough when AI raises deeper questions about identity, competence, fairness, and trust. Knowledge is not enough when knowledge expires faster than the training catalog can be updated. Ability is not only individual when the real constraint may be data quality, workflow design, governance, psychological safety, or leadership expectations. And reinforcement may even become problematic if we reinforce yesterday’s AI practice in a world where tomorrow’s better practice is already possible.

I will therefore suggest replacing ADKAR with a different logic:

• Create direction
• Make the work visible
• Experiment close to practice
• Build feedback loops, and
• Develop judgment.

Admittedly, CMEBD is not quite as catchy as ADKAR, and it will probably not become a global certification industry anytime soon. But I believe it is far more useful.

Because when everyone can produce more, faster, the real bottleneck becomes the ability to choose well. That requires direction. When AI adoption happens inside the work before it shows up in the program, leadership must make the actual work visible. When the most valuable use cases emerge where customers, employees, systems, and constraints actually meet, experimentation has to happen close to practice. When no organization will get this right in one attempt, feedback loops become more important than rollout plans. And when AI can produce confident, polished outputs at scale, human judgment becomes more important, not less.

That, to me, is the shift from change management to changefulness.

It is not a rejection of everything we know about change. People still need meaning, clarity, involvement, and support. But it is a rejection of the comforting idea that change can usually be managed as a neat movement from one stable state to another.

Organizations are already moving. AI simply makes that more visible, more urgent, and more consequential.

The leadership task is not to freeze organizations before changing them. It is to help them stay in motion without losing direction, judgment, or humanity.