A document that nobody finds is functionally the same as a document that does not exist. I wrote that line in the previous post almost in passing, but it has stayed with me. It has cost me time and effort over the years, and not just me but everyone around me in the organisations I have worked in.

A new finance manager spends two weeks writing a policy that already exists, because the original is filed under a different folder and nobody told them about it. A team reinvents a reconciliation template that a former colleague built three years ago and saved on a shared drive nobody looks at. A senior reviewer queries a number, gets an answer that contradicts a documented rationale from the previous quarter, and neither party knows the document is there.

The same thing is happening in each case. The knowledge exists. It is just not findable at the moment it is needed. So someone redoes it, or works without it, or makes a decision that contradicts it.

The hidden cost is the slow erosion of trust in the knowledge base itself. After enough people fail to find what they were looking for, they stop looking. They ask a colleague instead, or they reconstruct the answer from scratch. The repository becomes a place where things are filed for compliance reasons, not a place where things are found. Once that shift happens, it is hard to reverse.

This is why I think the AI-augmented version of knowledge enablement is genuinely promising, in a way the SharePoint generation was not. The right question is no longer “where is the document” but “what do we know about this.” When the system can answer the second question, the first one stops mattering. The document is still there, still version-controlled, still auditable. It is just not the thing the user has to reach for.

Whether organisations get there is mostly a question of whether they treat findability as a real cost. Most still treat it as a soft one.

— YS Lim