Reckoning

What I labelled as anxiety in AI transformation turned out to be grief. That realization changes everything about how we respond.

March 2026

What Prompted This

What prompted me to write this was a realization I kept circling back to in my own work. What I had been labeling as anxiety in the organizations I worked with, and if I am honest, in myself, was something else entirely.

It was grief.

The old world of work was dying. And nobody was naming it as a loss. We were treating a goodbye as a management problem.

Once I saw it that way, a lot of things that had puzzled me about AI transformation began to make sense. The resistance that did not respond to better communication. The disengagement that survived training programs intact. The leaders who said all the right things and still could not shift the mood in the room. None of it was

irrational. It was the behavior of people who were losing something real and had not been given permission to say so.

Where This View Comes From

I have spent years helping organizations navigate change. I know the frameworks. I have trained teams in them, built programs around them, and watched them work. For a period I also taught change management and organizational development at postgraduate level, including to MBA students who were, in many cases, living

through transformation in real time at the organizations they worked for.

It was in the classroom that I first noticed something shifting. Not in a single moment, but gradually. The frameworks I was teaching felt increasingly like maps drawn for a country that had quietly changed its borders. Students were bringing questions the curriculum did not have good answers for. The gap between the theory

and their lived experience kept widening, and I found myself improvising more and codifying less.

That instinct has been confirmed repeatedly in my practice work since, particularly in healthcare and technology organizations where AI is not a future consideration but a present operational reality. This paper names three things that are genuinely breaking down in AI transformation right now. Not to be provocative, but because HR and Operations leaders deserve an honest account of what they are actually dealing with, and what a more useful response looks like.

Breakdown One: Methodology Built for a World That No Longer Exists

The dominant change frameworks: ADKAR, Kotter, Prosci: share a foundational assumption: change is episodic. There is a current state. A desired future state. A gap to be managed. You move people across the bridge, embed the new behaviors, and the organization stabilizes.

This assumption is now structurally false.

AI transformation does not have a future state. By the time you have stabilized adoption of one capability, three more have emerged. The burning platform is not a metaphor: it is the permanent condition. Organizations are not crossing a bridge; they are learning to live on one.

Episodic change tools applied to climatic change produce episodic results. I have worked with technology organizations that completed well-structured, well-resourced AI transformation programs, closed them out with positive adoption metrics, and found themselves six months later no more adaptive than when they

started. The program was designed to end. The disruption was not.

What is needed instead is the development of ongoing change metabolism: the organizational and individual capacity to process disruption continuously, without waiting for stability that is not coming. That is a different goal than completing a program, and it requires a different kind of investment.

The implication for HR and Operations leaders is significant. The question is no longer "how do we manage this transition?" It is "how do we build an organization that can keep moving when the ground never stops shifting?"

Breakdown Two: An Identity Crisis Nobody Is Naming

The most consistent thing I observe in AI transformation: across sectors, levels, and geographies: is not resistance to the technology. It is a quiet, largely unspoken crisis of identity.

People derive meaning, status, and self-worth from what they are good at. When AI begins to perform those tasks: sometimes better, always faster: the emotional experience is not primarily "this is inefficient" or "I need training". It is closer to: who am I if this does what I do?

Standard change management addresses awareness and willingness. It does not address identity. And organizations that skip the identity layer find that even technically proficient adopters remain disengaged, resentful, or performatively compliant: going through the motions without genuine investment.

The professionals who navigate AI transformation best are not the most technically literate. They are the ones who have a robust sense of what they contribute beyond the tasks they perform. They have done identity work, consciously or not: and it shows in their resilience and adaptability.

This has a practical implication: before you invest in capability programs, ask whether people have the psychological foundation to use new capabilities without feeling diminished by them. If they do not, the training spend will underperform. There is also a grief layer here that organizations consistently underestimate. Even

when nobody loses their job, significant role change involves real loss: of mastery, of routine, of the collegial relationships built around how work used to flow. Organization's that do not create genuine space for this grief find it hardens into cynicism. The mourning does not disappear; it just becomes invisible and corrosive.

Breakdown Three: The Collapse of the "This Will Help You"

There is a particular kind of organizational dishonesty that has become pervasive in AI rollouts: the gap between what leadership communicates and what employees actually believe. The standard change communications playbook says: tell people what is changing, why it is good for them, and what support is available. In a trust-rich environment, this works reasonably well.

We are not in a trust-rich environment.

After years of restructuring, efficiency programs, and workforce reductions dressed in the language of "investment" and "opportunity", many employees have a finely calibrated radar for when they are being managed rather than told the truth. When leaders say "AI will free you up to focus on higher-value work", a significant proportion of the workforce hears "there will be fewer of you".

Sometimes that is accurate. Often the organization genuinely does not know yet. But the communications strategy proceeds as if certainty and goodwill exist, when neither does.

This pattern is particularly visible in healthcare and public sector organizations, where I have seen it play out repeatedly. Leaders in town halls and team briefings say the right things about people being the organizations greatest asset and AI being a tool to support rather than replace. Then they make decisions, communicate

timelines, and allocate resources in ways that contradict every word. Staff notice. They always notice. And each contradiction quietly withdraws another unit of trust that no communications campaign can replenish.

What rebuilds trust in this environment is not better messaging. It is better honesty. Leaders who are willing to say what they do not know, acknowledge what is genuinely uncertain, and demonstrate through their behavior rather than their words that people are not simply resources to be optimized.

This requires a different kind of leadership conversation. One that most organizations are not currently having, and one that most leaders have not been developed to lead.

What a More Useful Response Looks Like

None of this means change management expertise is obsolete. It means the center of gravity needs to shift: from process management to human system work. In practice, that means three things:

- Diagnostic depth before program design. Understanding where identity threat is highest, where trust is lowest, and who the informal leaders are whose stance will make or break adoption: before a single communication goes out.

- Leadership enablement, not just sponsorship. Leaders need to be able to have honest, human conversations about what is changing and why: not just to deliver a change narrative. That is a skill that can be built, but it requires genuine investment, not a briefing deck.

- Building change metabolism, not just managing change events. The organizations that will fare best over the next decade are those that develop genuine adaptive capacity in their people: the ability to process disruption, find meaning, and keep moving. That is a different goal than completing a transformation program, and it requires a different kind of sustained work.

A Final Observation

The organizations that are navigating AI transformation well right now are not the ones who got the technology right. They are the ones who got people right while the technology was happening around them.