Institutions produce enormous quantities of documents. Policies describe what should happen, procedures specify how tasks should be performed, databases preserve records, reports explain what has happened, and guidance attempts to translate accumulated experience into instructions that others can follow. This documentary infrastructure is indispensable to institutional continuity. Yet anyone who has spent enough time inside a complex organisation eventually encounters a more uncomfortable reality: some of the knowledge that makes the institution work may exist nowhere in its formal records.
An experienced employee knows that a particular type of case almost always becomes difficult when a seemingly minor condition appears. A programme manager understands why a procedure that looks inefficient on paper was introduced after an earlier failure. A technical specialist can recognise from an unusual pattern that a system is beginning to behave incorrectly long before an automated alert appears. A civil servant knows which organisation needs to be contacted when an issue falls between formal jurisdictions, and perhaps even which person inside that organisation will understand the problem. None of this knowledge necessarily appears in a manual. Some of it may never have been written down at all.
We often call this tacit knowledge, although the term covers several different things. Sometimes the knowledge could be documented but simply has not been. Sometimes people know how to do something but would struggle to explain exactly how they recognise the relevant cues. Sometimes knowledge depends so heavily on context that turning it into a general rule would remove much of what makes it useful. And sometimes what appears to be individual intuition is actually the result of years of accumulated exposure to recurring institutional situations. Experience can teach people distinctions that the formal organisation has never learned how to express.
This creates an important gap between the institution described in its documentation and the institution that actually operates. A procedure may say what should happen under normal conditions, while experienced staff know what to do when reality refuses to behave normally. An organisational chart may describe formal communication channels, while employees rely on informal relationships to resolve problems that cross administrative boundaries. Official guidance may provide general principles, while practitioners develop judgement about how those principles should be interpreted in unusual circumstances.
None of this means that informal knowledge is always superior to formal knowledge. Experience can preserve bad habits as easily as good judgement. Informal workarounds may conceal structural problems that should be corrected rather than perpetuated. Long-serving employees may defend practices whose original justification no longer exists. Tacit knowledge therefore should not be romanticised as a hidden reservoir of wisdom. But neither should it be ignored simply because it is difficult to see. Knowledge does not become institutionally unimportant merely because the institution has failed to document it.
The vulnerability becomes obvious when experienced people leave. Imagine someone who has worked in the same public organisation for twenty-five years. During that time, they have encountered policy changes, reorganisations, failed initiatives, unusual cases, crises and countless small operational problems. They know why certain processes exist, which exceptions matter, which warning signs deserve attention and which relationships allow difficult problems to be resolved. Then they retire. Their formal responsibilities can be transferred. Their files can be archived. Their successor can receive a handover document. But how much of twenty-five years of accumulated institutional experience has actually moved with those things?
Often, nobody knows until something happens that the departing person would have recognised. A problem that once took an hour to solve now takes several days. A team repeats an experiment that failed years earlier because nobody remembers the failure. A seemingly harmless procedural change removes a safeguard whose purpose had been forgotten. A relationship with another organisation deteriorates because the informal history behind it was never recorded. The institution discovers what it knew only after it has forgotten it.
This is one reason personnel turnover can affect institutions in ways that headcounts and organisational charts cannot capture. Replacing a person restores a position, but it does not automatically restore the knowledge that had accumulated around that position. Two employees can have the same formal job description while possessing very different understandings of how the institution actually works. The loss is therefore not simply human capital in an abstract sense; it can be a loss of institutional memory embedded in experience.
The obvious response is to document more, and often that is exactly the right response. Important decisions should have records. Lessons from evaluations should be preserved. Critical procedures should not depend entirely on one person’s memory. Institutions should identify knowledge that would create serious vulnerability if its current holders disappeared. But documentation has limits. Not every useful insight can be anticipated in advance, and attempting to record everything can produce enormous repositories that nobody can realistically navigate. The objective cannot be to convert every fragment of human experience into text.
A more realistic approach begins by recognising that different kinds of knowledge require different forms of preservation. Some knowledge belongs in formal documentation. Some can be transferred through mentoring, shadowing or structured handovers. Some requires communities of practice in which experienced people can transmit judgement rather than merely instructions. Some can be preserved through routines and organisational arrangements. And some may remain inseparable from human experience, requiring institutions to maintain enough continuity and overlap between generations for learning to pass from one to another.
This has become even more important in an era of digital knowledge management and artificial intelligence. New tools can make written institutional knowledge dramatically easier to search, retrieve and connect. They may help organisations rediscover reports, identify relevant precedents or make large documentary archives practically usable for the first time. But better access to what has been recorded does not solve the problem of what was never recorded. A sophisticated system cannot retrieve an institutional lesson that disappeared with the only people who understood it.
The deeper challenge is therefore not simply to build larger repositories. It is to understand which parts of institutional capability depend on knowledge that remains fragile because it is concentrated in experience, relationships or practice. Institutions need to know where that fragility matters most and create deliberate ways of reducing it without pretending that every valuable form of understanding can be fully codified.
The most important knowledge in an institution may never have been written down, but that does not make its disappearance inevitable. Institutions can learn to recognise the knowledge carried through experience, preserve what can usefully be preserved, create conditions for what cannot be easily documented to travel between people, and avoid discovering too late that an apparently routine departure has taken part of the institution’s capacity with it.