Seven traits separate KM contenders & pretenders

Knowledge management encompasses a broad range of ideas and issues. It's a concept that touches on practically every aspect of an organization, regardless of size, age, location and type of business. It is difficult to grasp because it brings seemingly contradictory disciplines together: on the one hand, the pragmatic, down-to-earth business world; on the other, the esoteric and abstract disciplines of philosophy, epistemology and social sciences. Thus, unless a more disciplined approach is taken to differentiate knowledge management from other IT trends, KM, too, will join the fossilized corporate fads.

Many IT services and product companies have jumped on the KM bandwagon by claiming to be knowledge management providers. Unfortunately, that has created further confusion about the concept. What is urgently needed is a list of functional capabilities that separate the pretenders from the contenders.

Seven KM characteristics

To create a knowledge infrastructure, the IS executive needs to integrate seven diverse yet related capabilities into a whole. The capabilities are critical in providing an effective knowledge support system. Each has the potential to develop into its own market segment and/or merge with other segments:

1. Business context

Capturing and monitoring user context incorporates a wide range of capabilities. The key is to enable the delivery of highest value information available by capturing all the characteristics that define a user's context, such as the following:

• general--industry, department, job function;

• security--access rights by repository to the document or object level;

• preferences--fields of interest, preferred format and frequency of delivery based on the type of information;

• process--the processes the user participates in, the groups or projects assigned to the user, the user's role in each process and the goals of the tasks performed in each role;

• information--the specific bits of information that proved useful as input for each task or decision, and the information that was created as output;

• device--PC, NC, PDA, cell phone, pager, etc.

Context must be constantly monitored because it is highly dynamic. Roles can change from one moment to the next. An emergency call from a customer can shift activity from one context to another without warning. Original work created on the desktop also represents a critical element of the real-time context of each user.

Capturing the meta data about users is a significant challenge; the following types of information are required to attack this issue at different levels:

User profiles must learn over time not only who a person is and what his/her interests and preferences are but be able to incorporate changes to the user's business context as they occur.

Information utility means that every time information is delivered to a user's desktop, the system must monitor what happens to that information and track whether it is read, copied, or re-purposed.

Context awareness is important for providing agent-based query facilities to support non-structured events such as writing a research paper or a business strategy. The technology must understand the user interface from a content perspective.

2. Knowledge brokering

Information repositories are continuously being updated, so the system requires the capability to broker the changing value of each bit of information in the face of continuous change on both the supply and demand side of the fence. A knowledge broker, therefore, performs the exchange between the context manager on the demand side and the content catalog on the supply side.

3. Content catalog

This is a key capability for organizing repositories based on content and not on file format. The key is that the cataloging is done in relation to the user base, and it is therefore being refined at all times. The benefit is that information silos are broken up and the resources of the enterprise can truly be applied to each individual information requirement.

4. Reference management

The three great challenges of searching a massive collection of repositories are:

• silence--not retrieving useful information due to misinterpretation of the query or misconstruction of the query itself,

• redundancy--retrieving the same item numerous times because it may reside in more than one repository,

• noise--retrieving many closely related but less useful sources of the same information.

Silence can be effectively attacked by technology that associates query items among many different users. For instance, if a user looks for three items, two of which are the same that others have searched on, the system should prompt the user that several items may be relevant to his/her needs, although they were not directly part of the query. Similarly, by referencing different manifestations of the same information and aggregating those sources, a reference management tool reduces the noise and redundancy delivered to the user. The result is faster assimilation and the use of more valuable information, which translates into better decisions that are made faster than ever before.

5. Decision process automation

Software in this category maps the business process of decision making through the application of graphical tools. That is a critical part of the KM world because by combining the core components of the decision process with the decision itself and the results of the decision, one has captured the context of the information and created knowledge from which another person can replicate or modify the process in the future.

6. Universal search and retrieval

A great deal of development activity is happening in this area, especially with respect to making search filters specific to vertical markets and functional areas. Natural language query and federated search capabilities will be de facto requirements.

7. Dynamic delivery

Dynamic delivery in a KM environment is the antithesis of the fire hose approach to access; the content provided is subject to refinement and filtering via content cataloging, reference engines and context managers. The only reason this is mentioned is because KM systems must be able to anticipate user needs as well as react to user requests, to provide the highest overall efficiency of knowledge support in the enterprise. The value add of dynamic delivery is the ability to support decisions when individuals don't even know to ask for help