Every company that discovers its oral tradition problem responds the same way: a documentation initiative.
A wiki gets stood up. A quarter gets dedicated. Everyone writes down their processes, there’s a spreadsheet tracking completion, and for one shining moment the company believes it has converted its folklore into infrastructure.
Come back eighteen months later. The wiki is a graveyard. Half the pages describe processes that no longer exist. The other half were written by people who no longer exist — at least not at this company. Nobody checks it before asking Sarah, because everybody learned quickly that Sarah is current and the wiki is archaeological. The initiative produced documents. It did not produce knowledge infrastructure. Those are different things, and the difference is the whole section.
—
I’ve been near this failure enough times to name its mechanics.
**Documentation freezes; reality flows.** A document is a photograph of a process on the day someone had time to write it. The process kept moving. Unless something forces the document to move too, every page begins dying the moment it’s published — and a wiki full of half-dead pages is worse than nothing, because it costs trust. One stale answer teaches a reader to never come back. You met this in Chapter 2 — the flowchart founder, the beautiful documentation, the five interpretations. Documented was never the same as true.
**Writing it down was nobody’s job, and keeping it true was nobody’s job squared.** The initiative assigned authorship once. It assigned maintenance never. The people who know the most are the busiest — the absorbers, naturally — so the documenting joins their invisible load for a quarter and then quietly stops. The wiki doesn’t fail dramatically. It just stops being fed.
**And the deep one: the most important knowledge resists writing.** Ask Sarah to document her process and she’ll write down the steps — the part a system could hold. What she can’t write down is the judgment: how she knows a client’s about to churn from the tone of one email. Which rules bend and which break. The forty exceptions she’s absorbed so completely they feel like instinct — the mask problem from Chapter 2, applied to expertise: she’s compensating so smoothly she can’t see her own compensations. The steps go in the wiki. The knowledge stays in Sarah. The company celebrates coverage it doesn’t have.
—
So if documentation isn’t the answer, what is?
The v1 answer from earlier in this manifesto still holds: not documentation for its own sake — visibility infrastructure. Systems that hold organizational knowledge in ways that don’t require a specific human to be present for the organization to function. The difference is *where the knowledge lives and works*:
A document about the pipeline is prose someone must find, read, and hope is current. Knowledge held in infrastructure is the CRM actually containing the deal’s real history — because the handoff payload from Chapter 8 made recording it part of the work, not homework after the work. The escalation rule isn’t a wiki page; it’s built into how tickets route. The definition of “qualified” isn’t a memo; it’s enforced by the form that won’t submit without it. Knowledge embedded in the flow of work stays current *because the work keeps it current*. Documents beside the work die; structure inside the work lives.
That’s the real test of any knowledge effort, and it fits in one question: **does the system still answer when the person is gone — and is the answer still true?** A wiki fails the second half. Oral tradition fails the first. Infrastructure passes both.
And Sarah? Watch what she becomes in a company that builds this. Not redundant — *promoted*, in the real sense. The steps live in systems now. What’s left of her seven years is the part that was never writable anyway: the judgment, the pattern-reading, the teaching. Which is what you were paying for all along, whether you knew it or not.
Oral tradition is not a revenue strategy. Neither is a wiki.
Reality infrastructure is. It’s also where this whole part of the manifesto has been pointing — because Part Four is about what happens when companies skip the diagnosis and buy the cure anyway.