What Cloud Transformation Actually Taught Me
9 min read
There is a version of this essay I could write that would be safe.
Technology alone is insufficient. People and culture matter. Organizations need alignment and change management. All of it is true, and almost none of it is useful as a guide to what actually happens inside large enterprises when a major transformation is underway. The conference version of these lessons has been written so many times that it no longer means anything to the people who most need to hear it.
I want to write the other version, because the same dynamics are playing out right now inside every large organization attempting AI transformation, and the leaders who recognize them will be the ones who get it right.
How the methodology was built
I spent nearly a decade at Amazon Web Services building the Enterprise Transformation practice at AWS Professional Services. I joined to build something that did not yet exist at scale: a practice dedicated to the organizational and human dimensions of cloud transformation. At the time, this was counter-cultural. AWS was a technology company, and the prevailing belief inside AWS and across the industry was that technology was the lever. Get the platform right, and adoption would follow.
I believed something different. I believed that culture, leadership alignment, and operating model were the real barriers between cloud investment and realized value. It was not a popular position at the start, but the evidence kept accumulating, and the methodology took shape through nearly a decade of collective learning. We built frameworks, took them into the field, watched what worked and what did not, and iterated continuously in real conditions with real organizations. The practice attracted an extraordinary community of practitioners, partners, and customers all working to help large complex organizations succeed at something they had not quite done before. What emerged was a body of work refined through hundreds of engagements and shaped by everyone in the room, not a theoretical construct that someone wrote down and asked the field to apply.
What the evidence showed
When cloud programs stalled, the assumed root cause depended on who you asked. Infrastructure teams pointed to security. Program leaders pointed to a lack of executive alignment. Executives pointed to a lack of technical expertise. Vendors got pointed at by everyone. And almost every customer believed they were operating in a uniquely complex situation, with technical, architectural, regulatory, and industry-specific challenges that meant the standard playbooks did not quite apply to them.
What was actually happening in most of these moments was something harder to name. Transformation is unsettling work. It changes the basis on which people understand their own value, asks them to give up the expertise they have spent careers building, and can make them look replaceable at exactly the moment they need to feel essential. When people are scared of becoming irrelevant, of no longer being recognized for who they are and what they bring, the response is rarely to surface that fear directly. The response is to dig in. To protect the boundary of one's own function. To explain why the problem lives somewhere else. And the energy that should be going to breaking down silos, learning together, and finding what is actually wrong gets spent instead on reinforcing the silos that were already creating friction in the first place. Self-preservation and fear-based decision-making are a rough combination, and they are present in every transformation I have ever been part of.
This is true at every level of the organization, including at the top. Executives are not exempt from the fear of irrelevance, and many of them carry it more heavily than the people below them because they have more visibly to lose. The leadership posture that transformation actually requires, leading through uncertainty, modeling curiosity, creating the conditions for honest contribution, is a rare combination on a good day, and transformation is not a good day. It is exactly the kind of pressure that makes the default leadership behaviors look more attractive than the ones the moment actually calls for.
Every transformation wave creates change, and change creates uncertainty about what people's roles, value, and futures will look like on the other side of it. That uncertainty is part of what drives the self-protective behavior described earlier, and it shows up in every major shift, not only this one. What is different about the current moment is the intensity and the omnipresence of the signal. AI is not a quiet, internal transformation conversation happening inside individual organizations. It is a daily public narrative, named explicitly as the reason for workforce reductions across the industry, and the people inside any given organization are absorbing that signal continuously whether or not anything has changed in their own building yet. The fear is no longer something a leader can address inside the boundaries of their own program. It is part of the air people are breathing, and pretending otherwise costs credibility quickly. The question becomes how to create the conditions for vulnerability, honesty, and innovation when there are real and visible reasons to be afraid, without dismissing the fear or amplifying it. That is the leadership challenge this moment demands, and it is one of the strongest arguments for why this wave requires a different kind of leadership than most organizations have developed.
The real reasons cloud programs stalled were harder to see, less convenient to name, and almost always organizational rather than technical. What I actually saw across hundreds of engagements was a different and more consistent pattern. Executive teams that appeared aligned were often quietly, structurally at odds. Not misaligned in ways they would publicly admit to, but with different priorities, different budget expectations, and different definitions of success that had never been resolved before the program launched. The transformation became a proxy battlefield for dynamics that predated it, and the program absorbed the cost of resolving conflicts that should have been worked through at the executive level long before any technology decision was made.
Many organizations radically underestimated the human and business impact of what they were undertaking. Leaders who approved what they understood to be a technology program discovered partway through that they were actually running a business transformation, one that required deep coordination across functions that had never worked this closely together, fundamental changes to how decisions were made and how work was done, and a clarity of future-state vision that most organizations had not developed before the migration began. The work was bigger than the budget, the timeline, and the org chart they had built for it, and the scramble to catch up was where many programs lost their footing.
Functions that felt bypassed tended to respond predictably. The security organization was a consistent example. Security teams were almost universally brought in too late, after major architectural decisions had been made without them, and they had every rational reason to resist a process they had not shaped. What looked like obstruction was a predictable response to a solvable design failure. The people who carry real accountability for risk need to be in the room at the beginning, and the programs that brought them in early ran differently from the programs that did not.
Then there is the leadership pattern that I think about more than any other, because it shows up in every transformation I have ever been part of and because it is showing up again right now in AI. Transformation requires something most leadership development has not prepared leaders for: the capacity to lead with genuine confidence in conditions of genuine uncertainty, to model curiosity, and to create environments where people feel safe bringing forward what they actually see. In programs that struggled, we often saw the opposite. Leaders who had been rewarded throughout their careers for having the answers continued to perform that certainty when they did not have it. The culture this created was not dishonest, exactly. It was silent. People learned that raising a concern would be characterized as resistance, and that the official narrative did not fully match the real one, and that the gap was not safe to name. Over time, the organization lost access to its most important resource, which is the honest intelligence of the people doing the work. Early warning signals went unheeded, the people closest to real problems were sidelined rather than heard, and the gap between stated values and lived reality grew until people simply stopped trusting the stated values.
The talent problem was almost always misread. Many organizations concluded that their people lacked cloud skills, which was partly true, and that training would close the gap, which mostly missed the point. We watched entire organizations complete certification programs and return to environments where they had no meaningful access to the tools they had just learned, where the conditions in the training environment bore no resemblance to what they actually encountered at work, and where months elapsed before they had any real opportunity to apply what they had learned. What people actually needed was relevant experience, continuous, integrated into real work, with access to real environments and the psychological safety to experiment and learn from mistakes in real time. Learning has to be part of the work, not a prerequisite for it.
The most damaging dynamic, and the one I would most want leaders today to take seriously, is the gap between stated values and lived reality. Many of the organizations I worked alongside had explicit cultural commitments to innovation, learning, candor, and psychological safety. In many of those same organizations, the lived experience of the people doing the work told a different story. People who raised early concerns, often accurate ones, were characterized as resistant rather than as contributors with valuable perspective. The behaviors that were publicly valued were privately penalized when they created friction. Over time, people learn the real rules, and they adjust their behavior accordingly, which means the organization loses access to the honest intelligence it most needs precisely when transformation pressure is highest. An organization that says it values learning but punishes failure is not learning. It is performing.
What the methodology addressed, and why integration matters
The insight that shaped the framework we built was that transformation is a system, and you cannot fix a system by working on its parts in isolation.
The Enterprise Transformation Framework was built around four integrated dimensions: business and strategy alignment, financial operating model, cloud operating model, and people and culture. These were not sequential phases or independent workstreams. They were concurrent forces that constrained and enabled one another. An organization could develop a technically excellent cloud operating model and plateau because its financial model still ran on legacy economics. It could invest deeply in people and change management and still stall because its leadership team had not resolved the strategic questions underneath. The dimensions were interconnected, and the methodology had to treat them that way, because nothing else worked.
This is not a framework designed to add complexity. It is a framework designed to reflect reality, and the reality is that the whole of an enterprise transformation is only greater than the sum of its parts when the parts are genuinely working together. Most transformations never get there, because the work of integration falls in the seams between functions and nobody owns the seams.
how the pattern is repeating for ai
Every one of these dynamics is active inside organizations attempting AI transformation today.
Executive misalignment is present and more consequential than it was in cloud, because AI changes operating models, workforce strategy, and competitive position at a level that requires genuine board-level alignment rather than the kind of executive sponsorship most organizations are comfortable providing.
The underestimation of human and business impact is widespread, and possibly more acute given the pace of change. The talent misreading is already repeating, with AI certification programs that build vocabulary without building judgment. The performance of certainty is endemic, in an environment where the honest answer to most questions about AI's organizational impact is genuinely uncertain and leaders are under intense pressure to appear as though it is not. The gap between stated values and lived reality is visible in organizations that announce AI as a strategic priority while the people closest to the work understand that what is actually being optimized for is short-term cost reduction and competitive optics.
The most consequential dynamic is one that did not have an analog in cloud at quite this scale. Organizations are removing the people whose judgment, institutional knowledge, and experiential expertise is exactly what AI systems depend on, in the name of the efficiency AI is supposed to create. The people who understand how work actually happens. The people who know where the data is unreliable and why. The people who carry the organizational memory that cannot be recreated from a prompt. This is not a workforce decision in the conventional sense. It is an AI infrastructure decision, and the consequences will surface slowly, long after the leaders who made them have moved on.
the Organizations that will lead
The organizations that will lead in the age of AI are not the ones moving fastest on the technology. They are the ones building the organizational infrastructure now, before the pressure is acute, before the shortcuts become structural, and before the people who could have told them the truth have been given enough evidence that telling the truth is not safe.
There is an important difference between this moment and the early days of cloud. The AI capabilities themselves are not settled. The tools are evolving rapidly, the competitive landscape is shifting continuously, and the regulatory environment is being written in real time. Organizations are not navigating a large complex change with a knowable endpoint. They are navigating change that keeps changing, at a pace of disruption that outstrips any organization's natural capacity to absorb it.
The organizations with transformation DNA are better positioned not because they have all the answers, because nobody does. They are better positioned because they have built the organizational capacity to navigate not having them. They can hold uncertainty without losing direction. They can learn continuously rather than episodically. They can close the gap between the culture an organization says it has and the one it is actually running.
That is the work. Not implementing the technology, because the technology will keep evolving regardless. Building the organization that can keep up with it. Building the leadership that can navigate it honestly. Building the culture where the truth is safe, learning is continuous, and transformation is not something that happens to the organization but something the organization knows how to do. That has always been the work, and it is more urgent now than it has ever been.
Amanda Rankin pioneered the organizational and human dimensions of enterprise transformation at AWS Professional Services at a time when the prevailing belief, inside AWS and across the industry, was that technology was the lever that mattered most. She championed and led the development of the practice and methodology that became embedded in AWS's most significant enterprise programs, including the Enterprise Transformation Framework and OCA 6-Point Framework, shaping cloud adoption across thousands of enterprises globally. She most recently led transformation from inside a large enterprise, applying these principles to a major data transformation program. She writes about enterprise transformation, organizational resilience, and what it means to lead through continuous change at amandarankin.me
