Most large organisations are exhausted by transformation. A Harvard Business Review article from early 2026 argues that the remedy is not another transformation programme. It is to get off the transformation treadmill entirely and build organisations that can adapt continuously.
The treadmill problem
The pattern is familiar. A new CEO or leadership team launches a transformation. It runs for two or three years, consumes enormous energy and produces some improvements. Then a new leader arrives, diagnoses fresh problems and launches another transformation. The organisation never recovers between cycles.
The cost is not just financial. Constant transformation burns out high performers, erodes trust in leadership and trains people to wait for the next initiative before acting. Momentum is replaced by cynicism.
Adaptive systems, not big bangs
The alternative is to design organisations that can absorb change as a normal operating condition. This means modular structures, clear decision rights, reusable technology platforms and a culture that rewards learning over compliance.
Adaptive systems do not eliminate the need for focused change. They make focused change possible without requiring a full organisational reboot. A pricing change, a new market entry or an AI deployment can happen within an existing operating rhythm rather than as a separate programme.
The organisations that adapt well tend to share a few traits. They keep teams small and accountable. They invest in reusable technology platforms. They document decisions so lessons can be reused. And they accept that some failure is part of learning, provided it is fast and cheap.
Incremental redesign
HBR’s argument for incremental redesign is not an argument for caution. It is an argument for making change cheaper and faster by reducing the overhead around it. When a system is adaptive, a small adjustment in one area does not require a cascade of steering committees, communication campaigns and training modules.
This is particularly relevant for AI adoption. Companies that treat AI as a transformation will struggle. Companies that treat it as an incremental capability — one that plugs into existing workflows and improves over time — will move faster.
What leaders should do
For executives, the practical step is to audit how much of the current change load is structural and how much is self-inflicted. If every improvement requires a programme office, the organisation is not adaptive. If change is constant but small, predictable and owned locally, it probably is.
The goal is not to stop changing. It is to stop treating change as a special event.