Institutions usually preserve the visible results of major reforms better than they preserve the experience of reform itself. Years after a reorganisation, it may be easy to reconstruct the new distribution of responsibilities, find the legislation that authorised the change, examine the organisational chart that emerged or identify the systems and procedures introduced during implementation. The institution can describe what changed and, in many cases, measure some of the consequences. Yet when a later generation asks a different question — how did the institution actually manage to change? — the answer can be much harder to recover. The negotiations that made the transition possible, the resistance that nearly stopped it, the temporary arrangements that allowed old and new systems to coexist, the sequence in which capabilities were built and the informal adaptations through which people made the new architecture workable may have largely disappeared from institutional memory.
This distinction between remembering the outcome of reform and remembering the process of transformation matters because institutional change is itself a form of experience. When an organisation moves from one configuration to another, it generates knowledge not only about the destination but about its own capacity to move. It discovers which dependencies become critical during transition, where resistance emerges, which actors can translate between old and new arrangements, how long apparently simple changes actually take and which temporary solutions prove necessary before a new system can function reliably. Much of this knowledge is highly specific to the institution because it reflects its history, authority structures, professional cultures, technological dependencies and relationships with the wider environment. If that experience disappears once the reform is complete, the institution may preserve the result of transformation while losing much of what the transformation taught it.
This can happen surprisingly easily because reform processes create strong incentives to document destinations rather than journeys. Formal decisions need to specify new responsibilities, implementation programmes need milestones and budgets, technical projects produce specifications, and evaluation tends eventually to ask whether the intended outcomes were achieved. These records are valuable, but they often privilege the architecture that emerges over the mechanisms through which the organisation managed to reach it. Once the new arrangement becomes normal, temporary structures disappear, project teams dissolve, external advisers leave and the people who negotiated difficult transitions return to other responsibilities. What was intensely visible during the period of change gradually becomes background history.
The resulting memory can therefore be strangely asymmetric. An institution may know exactly when a new organisational model was introduced while remembering little about why one part of the transition succeeded and another struggled. It may retain the final governance framework while losing the sequence of compromises through which different actors accepted it. It may preserve the technical design of a new platform while forgetting that implementation became viable only after workflows were changed, informal support networks emerged or experienced employees translated between technological and operational perspectives. The final state remains legible, while the pathway that produced it becomes increasingly difficult to reconstruct.
This matters whenever the institution needs to change again. Without a memory of transformation, every major reform risks being treated as though institutional change itself were an unfamiliar problem. New teams rediscover the importance of transition capacity, encounter predictable forms of resistance as though they were surprising, underestimate dependencies that earlier reforms had already exposed and reconstruct coordination mechanisms that previous generations developed under similar pressure. The institution may therefore become experienced at having been reformed without becoming experienced at reforming itself. Its history contains multiple transformations, but those transformations do not necessarily accumulate into a usable capability for future change.
A memory of change has to preserve more than a narrative of success. Indeed, some of its most valuable material may concern moments when the intended transformation did not proceed according to plan. A reform that stalled can reveal where authority was insufficient, where incentives conflicted or where the proposed architecture depended on capabilities that did not yet exist. A transition that required repeated adjustment may expose hidden relationships between parts of the institution. An implementation that succeeded only after its original timetable was abandoned may reveal something important about the pace at which organisational practices can genuinely change. If institutional memory preserves only the official account of what was ultimately achieved, much of this evidence can disappear behind the apparent coherence of the final outcome.
This is one reason why retrospective accounts written after successful reforms can be misleading even when factually accurate. Once the destination is known, the pathway can appear more orderly than it actually was. Decisions that were uncertain at the time acquire an air of inevitability, abandoned alternatives disappear from view and temporary arrangements are treated as incidental rather than as possible conditions of success. The institution remembers a sequence whose meaning has been reorganised by hindsight. Preserving the memory of transformation therefore requires attention to uncertainty as it was experienced, including which choices were genuinely open, what actors believed at different moments and which developments could not have been confidently predicted in advance.
Informal knowledge is particularly vulnerable. Transformation often depends on people who bridge organisational boundaries, interpret ambiguous decisions, resolve practical conflicts or maintain continuity while formal structures are changing. Their contribution may never appear clearly in official reform documentation because it is relational rather than procedural. Yet removing these elements from the historical account can leave future reformers with a distorted picture in which formal authority appears to have produced change almost by itself. The institution remembers the decision but forgets the social and cognitive infrastructure through which the decision became operational.
Digital transformation offers many examples of this problem because the technical artefacts of change are usually easier to preserve than the organisational learning surrounding them. An institution may retain extensive documentation for a platform, data architecture or automated process while gradually losing knowledge of how staff adapted their work, which assumptions proved unrealistic, where human judgement remained necessary and why particular compromises were made between technological possibility and institutional reality. When another technological transformation begins, the archived system may provide abundant information about what was built while providing much less insight into how the organisation learned to build and absorb it.
Artificial intelligence makes this distinction still more consequential. Institutions introducing AI-supported processes will generate experience not only about particular systems but about how authority, expertise, contestability and human judgement are renegotiated when new cognitive technologies enter existing workflows. If the institution records only which tools were eventually adopted and how they performed, it may lose the more difficult knowledge generated through adaptation: where professional resistance revealed legitimate concerns, which oversight arrangements became necessary, how confidence in automated outputs evolved and which organisational interfaces had to change before the technology became usable. The durable asset may therefore lie partly in the transformation experience itself rather than only in the technology that survives it.
Preserving a memory of change does not require every reform to become an exhaustive historical archive. The objective is not comprehensive documentation but cognitive continuity. Institutions need enough evidence of transformation processes to reconstruct how change actually occurred: important dependencies, turning points, abandoned approaches, temporary arrangements, sources of resistance, unexpected enabling conditions and the reasoning behind significant adaptations. Some of this can be captured through project records and evaluation, some through reflective accounts from participants and some through deliberate efforts to preserve the experience of teams before they disperse. What matters is that the pathway does not disappear entirely behind the destination.
Nor should this memory become a formula for repeating the past. A mechanism that enabled one transformation may fail under different conditions, and lessons drawn from earlier reforms can become constraints if they are treated as universal rules. Remembering change is valuable because it provides evidence, not because it provides instructions. The institution gains a richer basis for asking why a previous transformation unfolded as it did, which conditions were specific to that moment and which patterns might be relevant again. The memory becomes useful when it supports interpretation rather than imitation.
This is also why a memory of transformation should remain distinct from the later task of learning how institutions change. Preserving the experience is a necessary condition for such learning, but it does not perform the learning automatically. Records of resistance, sequencing, temporary structures and adaptation still need to be interpreted before the institution can infer anything reliable about its own transformation capability. Memory of change creates the evidence from which institutional learning about change can later become possible; it should not be confused with the lessons eventually drawn from that evidence.
An institution that can imagine being different has acquired an important prospective capability, but imagination alone cannot make institutional evolution cumulative. Transformation also needs a retrospective capability through which earlier attempts to change remain available as experience rather than disappearing once their outcomes become normal. Without it, each generation can inherit a transformed institution while knowing surprisingly little about the institutional work through which transformation became possible.
Institutions therefore need to remember more than what reform produced. They need enough memory of how change actually happened to preserve the uncertainty, resistance, sequencing, adaptation and enabling conditions through which one institutional configuration became another. Otherwise an institution may transform repeatedly without becoming progressively more knowledgeable about transformation itself, carrying the visible products of its history into the future while continually losing the experience of how it learned to change.