by Karl Wiig,
Knowledge Research Institute,
Inc. (edited by Limited Edition staff)
The following describes one of 16 knowledge management
"building blocks" that Karl has identified for a forthcoming
book. This is a summary; the full
text is available on the Society Web
(membership required).
After knowledge has been acquired or otherwise "sourced,"
it usually must be changed – transformed – to become a valuable
asset or to facilitate an application. Some transformation functions are
institutionalized and require formal (explicit) methods, but the vast
majority are performed on an ad hoc and informal basis as part of individual
employees’ regular workday. Four kinds formal transformations are
discussed below:
Compiling and organizing;
Embedding and tool-making;
Disseminating through people and automated methods;
Creating new products and services..
Compiling and organizing
As knowledge is acquired, it can be compiled manually or automatically
into a repository, which involves selecting, verifying and validating
knowledge for appropriateness, correctness, and completeness. Subject
matter experts (SMEs) from the business area need to be part of the validation.
Collected knowledge destined for repositories (such
as knowledge bases, data bases), must be organized according to an established
ontology (categorization of "what is"). These tasks can best
be performed by individuals with understanding of the target topic areas
and knowledge organizational principles (e.g., Library Sciences, Epistemology).
Embedding and tool-making
Knowledge can be transformed in many ways. It can be embedded into intermediate
products, new customer products, or educational programs or used to create
personal or shared work tools for decision making. This is commonly done
by insurance underwriters or securities portfolio managers.
Knowledge can be transformed at the same time as it
is compiled and organized but the type of transformation varies with the
application. For example:
Knowledge destined for case-based reasoning application
(such as a help-desk problem solving application) must be transformed
using the "crisp" or "fuzzy" rules that support
rule-based reasoning and must be edited into the case format.
Knowledge destined for training and education programs
must be reorganized to fit the teaching formats and delivery vehicles
(e.g., one-to-many personal knowledge transfer or interactive multi-media
computer-based delivery).
If the source knowledge exists in numeric forms
(or in advanced cases, as natural language), a selective transformation
can take place by using automated knowledge discovery in data bases
(KDD). This method still requires considerable manual interaction.
When the knowledge will be used for education in
general principles, anchored in concrete examples, the source information
must be analyzed to identify the underlying meanings, using hermeneutic
methods. This is particularly the case when knowledge is prepared for
teaching schemas and scripts, with concrete examples to let recipients
internalize and build "systematic," "pragmatic,"
and "automatic" knowledge and acquire the understanding of
how to apply it in practical situations. To facilitate knowledge construction
and acquisition, the education may also use conceptual maps.
Disseminating through people
For a knowledge-vigilant environment to function, individuals must be
motivated to share knowledge, work together in collaborative settings,
and rely upon each other to secure delivery of quality work. Keys to success
are recognition targeted to specific on-the-job situations and sufficient
time for information exchange. Examples include:
Apprenticing and "shadowing" are powerful
practices when the need is to transfer deep and comprehensive bodies
of knowledge
Collaborative teamwork is useful for knowledge transfer
between a group of experts and practitioners or knowledge "pooling"
among experts with narrow specialties.
Expert networks are used to make the knowledge assets
possessed by individuals available to people at the point-of-action
(PoA). Expert networks are useful in situations where the person at
the PoA has time to consult with an outside party without degrading
the quality of delivered work.
Instructor-led educational and training programs
have been the standard approach to provide knowledge and understanding.
They are expensive and require comprehensive preparation but can in
many situations not be replaced.
Dissemination by automated methods
Automated methods of disseminating knowledge are increasingly common in
customer service applications. The technologies come in a wide range and
include case-based reasoning (CBR), fuzzy or true-false logic, expert
systems, neural nets, and natural language processing. Examples of fully
automatic KBS applications include:
