Global Governance & Public Policy
Current models tolerate digitization because it preserves its logic. They struggle when confronted with changes that alter the logic itself. So, changing or not – today’s risk is not that organisations fail immediately. The risk is subtler: that legacy processes are automated without questioning those processes are fit for a world of distributed intelligence. Automation without questioning yesterday’s harden structures into tomorrow’s constraints. Should Hierarchies Be Abandoned? An AI-augmented organisation would not necessarily eliminate hierarchy, but supplement it with continuous feedback structures. Instead of periodic review, oversight becomes embedded. Authority would remain, but it would operate in dialogue with real-time intelligence rather than after-the-fact reports. Why Old Processes Persist There are many reasons why legacy systems survive. First, incentives don’t reward architectural redesign. Chief executives operate on quarterly results. Politicians operate within electoral cycles. Organisational transformation, by contrast, unfolds over years and carries execution risk. Second, governance structures are constrained by law. Public administration’s legitimacy depends on procedural consistency. Private corporations have regulatory expectations embedded in traditional reporting models. Third, human psychology favours familiar lines of control. Delegating operational judgment to machines challenges deeply held assumptions about authority and expertise. Finally, there is an epistemic gap: alternatives are still emerging. We lack tested models of post-industrial governance at scale. It is therefore understandable that organisations hesitate. But demographic pressures, fiscal constraints, and geopolitical competition narrow today’s hesitation window. Automating routine administrative functions is not merely efficient; it is economically necessary. The question is whether automation reinforces old procedures or enables new ones. What Could Reinvention Look Like? An organisation designed for integrated AI would likely begin not with org charts, but with decision flows. Some decisions would remain human. Others would be machine-assisted but subject to transparent audit trails. Governance becomes continuous assurance instead of periodic supervision. Boards receive structured, real-time indicators
of system performance and anomaly detection. Oversight committees evolve from reactive bodies into architectural stewards, responsible for the integrity of decision systems rather than merely their outputs. Agentic AI, properly designed, serves as a simplifier. It could identify redundant approval layers, simulate alternative process pathways, and test the impact of structural changes before they are implemented. In this way, applications would assist in redesigning the very processes they operate within. The danger is equally clear. If agentic systems are trained exclusively on historical workflows, they will encode and perpetuate those workflows. Automation without conceptual rethinking risks cementing the past in code. Leaders must ask not only how a process can be automated, but whether it should exist in its current form at all. One Organisation at a Time or Systemic Shift? Unless caused by large external events e.g., war, financial crashes, epidemics… (societal or organisational) change and/ or reinvention rarely occurs simultaneously across society. Since change objective is not disruption for its own sake. It is resilience and so proceeds incrementally. Yet local redesign can aggregate into systemic change. Financial reporting standards were not adopted globally overnight; they diffused through practice and necessity. The same may occur with AI-adapted governance models. If It Works, Why Change It? It is reasonable to ask whether transformation is necessary when current systems function “adequately.” But adequacy is context dependent. Organisations designed for slower information cycles and abundant labour may not remain adequate in an environment defined by instantaneous data flows and demographic contraction. The question, therefore, is if leaders are willing to examine foundational assumptions rather than merely accelerate inherited structures. AI offers extraordinary capability. Whether it becomes a tool of resilience or a mechanism of rigidification depends less on model performance than on organisational imagination. Acceleration is technologically straightforward. Reinvention is politically and organizationally demanding.
Choosing Structure: How Leadership Can Act Deliberately
Choosing structure means consciously deciding which organisational assumptions remain valid and which no longer correspond to reality. That choice can be made, but it requires new forms of inquiry at the highest level. DAVOS CLUB 25
Powered by FlippingBook