— Has AI Introduced a New Layer Called Knowledge Routing? —
In my previous article, I questioned a growing assumption that I hear more and more often:
"Now that AI can generate documentation, we no longer need static documentation."
My concern was never about AI's capabilities.
It was about the assumption behind that decision.
Abandoning static documentation is, in effect, a bet on a particular future:
"Humans will never need to read static documentation again."
This time, I'd like to look at that question from a more theoretical perspective.
What Knowledge Management Has Been Explaining
Knowledge Management (KM) studies how organizations create, preserve, share, and utilize knowledge.
One of its best-known theoretical foundations is the SECI model.
SECI describes how Tacit Knowledge and Explicit Knowledge continuously transform into one another, creating an upward spiral of organizational learning.
Even today, I believe it remains an elegant and highly relevant model.
The arrival of AI does not invalidate SECI.
So What Has AI Actually Changed?
AI appears to generate knowledge.
In reality, it reads existing knowledge sources, combines them, and presents the result in a form appropriate for the current context.
The important change is therefore not the definition of knowledge itself.
The change lies in how knowledge reaches its destination.
Knowledge now routinely passes through AI before reaching another human.
That simply wasn't part of the environment when SECI was originally proposed.
What SECI Doesn't Try to Describe
Looking back at SECI, I noticed something interesting.
SECI explains how knowledge is transformed.
Tacit becomes Explicit.
Explicit becomes Tacit again.
This transformation forms the famous knowledge-creation spiral.
However, SECI is not primarily concerned with another question:
Who carries knowledge to whom?
Or more precisely:
Through what path does knowledge travel?
Human communication has always existed, of course.
But AI has now become an active participant in that communication process.
Perhaps AI Added Another Layer
I don't think AI replaces Knowledge Management.
Instead, it seems to have introduced another architectural layer beneath it.
Knowledge Management explains how knowledge is created and maintained.
The AI era forces us to think about something different:
- Who consumes knowledge?
- Who should receive it?
- Which knowledge source should be selected?
- Should the answer come directly from static knowledge, or should it be dynamically generated?
These are no longer questions about knowledge itself.
They are questions about knowledge paths.
Who Is the Knowledge Router?
This naturally raises another question.
Who—or what—is the Knowledge Router?
Is it AI?
A human expert?
A knowledge base?
Or perhaps several of them working together?
Knowledge is no longer simply stored and retrieved.
It now passes through multiple participants before reaching its destination.
That behavior seems different enough to deserve its own perspective.
Toward Knowledge Routing
If this perspective is useful, it also changes how we think about documentation.
Static and Dynamic are no longer merely categories of documents.
They become characteristics of different routing strategies.
A Dynamic Document may not be a document in the traditional sense at all.
It may simply be a view generated as the result of a routing process.
This is still just a hypothesis.
But at least to me, AI doesn't seem to replace Knowledge Management.
It seems to introduce another architectural layer beneath it.
For now, I'd like to call that layer Knowledge Routing.