From e-government to AI transformation: what four decades of institutional technology change teach us about governance.

Every major technology wave of the past four decades has produced the same split outcome. A portion of organizations embedded the new technology into their core operating models, restructured how work was actually done, and carried a durable competitive advantage into the next cycle. A larger portion added the same technology as a layer on top of existing processes, declared themselves modernized, and then spent years managing the gap between what they expected the technology to deliver and what it actually delivered. The pattern is consistent. The underlying cause is almost always the same.

The organizations that struggle are not struggling because the technology failed them. They are struggling because they adopted the technology without restructuring the human operating model alongside it. Governance came after the fact, if it came at all. The integration question was treated as a technical question rather than a strategic one. And the window for getting it right the first time closed before they recognized it had opened.

I have watched this pattern repeat across multiple technology cycles, in multiple institutional contexts, across multiple geographies. What is unfolding with AI today is not a new phenomenon. It is the latest instance of a pattern that has been observable for a long time.

The Institutional Evidence

In the early years of the internet, when tools such as document management, electronic signatures, GIS, ERP, IVR, and workflow automation represented an entirely new generation of capabilities, I was part of an expert advisory team that contributed to the development of the e-government strategy for Dubai Municipality. The work involved restructuring both the internal processes of a major government institution and its external service delivery to citizens, building the operational foundation that shapes how Dubai Municipality functions today. Digital transformation was not yet a term in common use. We were simply trying to build institutional infrastructure to serve a rapidly growing population with the speed, accuracy, and efficiency made possible by modern technology for the first time, and we were doing it without an established playbook for what institutional transformation at that scale actually required.

What we learned during that time has stayed with me through every technology cycle since. The organizations that successfully embedded these new capabilities into their operating model shared a specific characteristic: they were willing to redesign how work got done, not just which tools were used to do it. The governance and workflow redesign challenges were inseparable from the technology adoption challenge. Organizations that treated them as separable almost always ended up with expensive systems that were underutilized, underperforming, or resisted by the people they were meant to support.

The parallel experience in the U.S. Department of Defense reinforced the same observation at a different institutional scale. As part of a DOD initiative to consolidate R&D laboratories across the military services, I was involved in institutional restructuring that required systems thinking, stakeholder alignment, and governance architecture to be applied simultaneously within one of the most complex organizational environments imaginable. I also led development of one of the first environmental decision support systems in the DOD, applying technology to governance and decision-making at a time when that combination had no precedent. The technology was not the central challenge. The governance architecture around it was.

In both institutional contexts, the pattern was the same: successfully integrating technology as an operational issue worked. When technology was approached as a procurement challenge and merely added as a necessary layer on existing processes, it accumulated structural debt that later required costly unwinding.

The Procurement Trap

The organizations struggling most with AI adoption today are not struggling because AI is inadequate. They are struggling because they have treated AI as a procurement decision and merely added it as a necessary layer on existing processes. The question driving most AI investment discussions at the leadership level is: which tools should we adopt? That question is understandable. It is also the wrong question to be leading with.

The question that actually determines the outcome is different: what does our business model need to become for AI to deliver its intended value?

Governance and workflow discipline are not the destination of AI adoption. They are the execution mechanisms that enable operational integration. They are how you get from a procurement decision to a structural capability. An organization that selects AI tools before it has answered the business model question will spend significant capital on tools its teams cannot use effectively, create governance risks it has not anticipated, and find itself restructuring a failed initiative rather than capturing the value it imagined.

McKinsey's 2026 State of AI survey reports that nearly nine in ten respondents use AI regularly in at least one business function, but only 6% are AI high performers, meaning they attribute an EBIT impact of 5% or more to AI and report significant value. The gap is not a technology problem. McKinsey's own analysis shows that nearly three-quarters of those high performers have fundamentally redesigned their workflows around AI, compared with one quarter of other respondents, and that they are twice as likely to report senior leaders who demonstrate commitment and defined processes to measure impact. The tools are present across the overwhelming majority of organizations. The organizational preconditions for those tools to deliver value are not.

The organizations that will carry a structural advantage from this AI cycle into the next one are those that are currently doing the harder work alongside the tool selection. They are asking what their business model needs to become. They are restructuring workflows before deploying tools into them. They are establishing governance frameworks that will scale with their AI ambitions rather than needing to be retrofitted after problems emerge. They are governing AI adoption from the beginning rather than after the first costly failure has made governance feel urgent.

Where AI Stops, and Judgment Begins

Every AI system will eventually face a decision it cannot make. The question is whether your governance structure knows that before it happens.

I have sat across the table from leadership teams discussing this very issue more than once. Most were entirely oblivious to the governance question. No one had asked it before the system was deployed, and few had thought through the soft, hidden implications or the liabilities they were quietly placing on their organizations by not asking it. The gap between what they believed they had approved and what they had actually authorized was significant. By the time it became visible, the cost was already running.

This is not a theoretical risk. It is a governance requirement and one of the most consequential design decisions your organization will make in this AI cycle. The organizations that get it right build governance structures that preserve and enhance human decision-making capacity at exactly the moments when it matters most. The organizations that get it wrong end up with AI systems processing consequential decisions without the oversight infrastructure to catch errors before they compound. That gap does not announce itself. It accumulates quietly, and it arrives faster than most leadership teams expect.

The organizations that govern AI well did not build sophisticated frameworks. They asked a simple question before deployment and answered it honestly: where does the human have to remain in the loop, and what happens if they are not there? They keep human judgment anchored at the critical decision points while allowing AI to do the amplification work around them. The organizations I have seen navigate this well share one characteristic: they defined that boundary before they deployed a single tool, not after the first failure made it feel urgent.

That balance is achievable. It does not happen by default. And the window for designing it correctly is open right now.

The urgency argument, stated plainly

Let me be direct about what is actually at stake.

Organizations adding AI as a layer on existing processes are not making a business-model decision. They are avoiding one. They are making a procurement decision for a legacy operating model that is already becoming obsolete, and with every tool they add, every workflow they automate without restructuring, and every governance question they defer, they are making that model harder to untangle and more expensive to rebuild. The structural debt accumulates quietly. By the time it becomes visible, the cost of unwinding it will exceed the cost of having done it correctly from the beginning.

The consequences of getting this wrong are not theoretical. An organization that deploys AI without first answering the business model question will find that its tools are optimizing for efficiency within a structure that needs to be fundamentally redesigned. An organization that treats governance as an afterthought will discover its liability exposure through an incident rather than through a framework. An organization that fails to define where human judgment must remain in the loop will gradually lose that boundary, without a single moment of decision, until the moment of consequence arrives and there is no governance structure in place to contain it.

The window for getting these decisions right in this cycle is narrowing. Not because the technology is moving too fast, but because organizational decisions that feel reversible early in an adoption cycle become structural over time. The patterns I observed across e-government implementation, DOD institutional restructuring, and digital transformation programs across multiple industries all confirm the same outcome: organizations that treated each technology wave as a business model and governance challenge emerged stronger and faster. The organizations that treated it as a procurement and implementation problem are still carrying the weight of that choice, often without fully understanding why they fell behind.

This is not a prediction. It is a pattern observation from someone who has been in the room across multiple cycles. The alarm is appropriate. The time to act on it is now.

The leaders who will look back on this cycle as their advantage started here

If you have read this far and recognized your organization in any part of what I have described, the most important next step is not another internal discussion about which tools to evaluate. It is an honest, structured assessment of where your organization actually stands before the next commitment is made.

That is precisely what the AI Governance, Transformation, and Operating Value Review is designed to do. It begins where the real questions live: not with your technology stack, but with your business model, your governance architecture, your operating workflows, and the human accountability structures that will determine whether your AI investment delivers value or compounds the problem. It produces a clear picture of where AI can create measurable operating value, where adoption is premature or poorly governed, and what the path forward should be before committing further capital. The work is structured, and the outcome is practical, implementable, and impactful.