1
Q
Data
A
Transactions, events, entities captured by firm’s Transaction Processing Systems. No context.
2
Q
Information
A
Data organized and summarized using categories such as time, place, type, etc.
3
Q
Knowledge
A
Discovery of rules, relationships, patterns that focus on the WHY, WHAT and HOW.
4
Q
Wisdom
A
the ETHICAL use of knowledge to solve organizational and societal problems.
5
Q
Dimensions of Knowledge-Knowledge Is An Asset
A
– It is Intangible
– The transformation of data into knowledge requires considerable resources.
– Knowledge is not subject to the Law of Diminishing Returns.
– Rather it experiences Network Effects and increases in value as more people use it.
6
Q
- Knowledge Is Situational
A
– Knowledge is conditional in that knowing when apply a procedure is just as important as how apply it.
7
Q
The KM (Knowledge Management
) Value Chain
A
consists of five phases
– Acquisition
– Discovery
– Storage
– Dissemination
– Application
* Value is added in each phase.
8
Q
The KM Value Chain Acquisition
A
- Until the 1990s a huge quantity of “knowledge” lay unused in databases, documents, email, presentations, etc.
- Know how also resided in people’s memories. These were the ”byproducts” of work and no one paid it much attention.
- In the 90s, companies began to realize that the knowledge that resided in these artifacts was a strategically valuable resource.
- Thus began the quest to collect the knowledge stored in these forgotten reservoirs.
9
Q
A
- Organizations are using progressively more esoteric technologies to discover hidden patterns and relationships.
- Neural Networks and Genetic Algorithms are being used to mine through these data and information deposits.
- Knowledge workers are also being equipped with advances Knowledge Work Systems to aid in the discovery and production of new Knowledge.
10
Q
Tacit knowledge
A
Personal, experience-based knowledge that is difficult to codify (e.g., intuition, skills).
11
Q
Dimensions of Knowledge
* Knowledge Has Different Forms
A
– Knowledge can be Tacit or Explicit
– Knowledge involves know how, craft and skill
– Knowledge is about the why in addition to the how and when.
12
Q
Dimensions of Knowledge -Knowledge Has Location
A
– Knowledge is a cognitive event involving mental models and maps. Thus it resides in us.
– Knowledge is situated and contextual. Thus it is not always easily transported to a different place or context.
13
Q
Dimensions of Knowledge- Knowledge Is Situational
A
– Knowledge is conditional in that knowing when apply a procedure is just as important as how apply it.
14
Q
The Knowledge Management Value Chain consists of five phases
A
– Acquisition
– Discovery
– Storage
– Dissemination
– Application
15
Q
The KM Value Chain Acquisition
A
- Until the 1990s a huge quantity of “knowledge” lay unused in databases, documents, email, presentations, etc.
- Know how also resided in people’s memories. These were the ”byproducts” of work and no one paid it much attention.
- In the 90s, companies began to realize that the knowledge that resided in these artifacts was a strategically valuable resource.
- Thus began the quest to collect the knowledge stored in these forgotten reservoirs.
16
Q
The KM Value Chain Discovery
A
- Organizations are using progressively more esoteric technologies to discover hidden patterns and relationships.
- Neural Networks and Genetic Algorithms are being used to mine through these data and information deposits.
- Knowledge workers are also being equipped with advances Knowledge Work Systems to aid in the discovery and production of new Knowledge.
17
Q
The KM Value Chain Storage
A
- Once organizations realize that this knowledge is a critical resource they are no longer willing to leave it “lying about” in databases, filing cabinets, email systems or personal computers.
- Neither do they want to leave it “walking about” trapped in the minds of employees who go home every afternoon.
- Now “Knowledge” is stored in knowledge warehouses. These are databases that can store files, documents, video, email, presentations, etc.
- Every item is “indexed” or “tagged” so that it can be found quickly. “indexing” also allows for discovery of relationships between items – say a presentation, a document and an email message by a past employee.
18
Q
The KM Value Chain Dissemination
A
- Once we have the knowledge “stored” we need some way to get it to the people who need it.
- Portals, search engines, groupware have been used to disseminate the vast amount of material stores in these knowledge repositories.
- Companies have also created collaborative “expert discovery” systems that allow users to discover a subject expert who may work somewhere else in the company.
- Unfortunately, knowledge workers have a new problem. That is how to zero in on the material they need when so much information is coming at them.
- This is known as the “drinking from a fire hose” problem.
19
Q
The KM Value Chain Knowledge Application
A
- Ultimately, KM is a means to and end. The end is the willingness of managers to use the knowledge to create new processes, products, ideas and markets.
- In order for KM to work Management must create a cultural environment where decisions are based on informed knowledge and not on guesswork.
20
Q
Types of KM Systems
A
-Enterprise Wide KM Systems
– Knowledge Work Systems
– Intelligent Techniques
21
Q
Enterprise Wide KM Systems
A
Enterprise KM systems are so named because of their enterprise wide impact. These systems support employees who may be dispersed in different geographies or in different functions.
- Enterprise systems are designed to handle structured, semi structured and tacit knowledge
22
Q
Enterprise Wide KM Systems Structured
A
Structured Knowledge Management Systems capture and store the explicit knowledge held in documents, reports, etc.
- The challenge in creating Structured KM system is developing an indexing scheme for each item so that it can be “tagged” “stored” and “retrieved” from a database.
- Documents fed into KPMG’s KWorld system are “indexed” based on 23 industry segments, 21 KPMG practice areas and 9 content types.
- Thus a document could end up in any one of 4,347 subsections in the KWorld document repository.
23
Q
Enterprise Wide KM Systems Unstructured
A
- Unstructured KM systems attempt to capture and store the vast amount of unstructured information held in folders, documents, memos, proposals, email messages, etc.
- The challenge here is building a database that can capture and store this information in a coherent fashion.
- Each item that is put into the system must be indexed or “tagged” in some meaningful way. Otherwise users would not be able to find it.
- Manually evaluating and tagging each document would be prohibitively expensive.
- Several tools claim to do “auto tagging”. They lower the cost of “tagging” but may not be as accurate as a human.
24
Q
Enterprise Wide KM Systems Knowledge Network Systems
A
- Despite the benefits of these structured and unstructured repositories much of the knowledge in an organization is stored in the minds and memories of its employees. It never gets onto paper!
- This leads to a huge amount of redundant effort as people constantly reinvent the wheel. IDC estimates that each company in the Fortune 500 waste US$ 60 Million a year.
- I suspect that some waste considerably more.
- Knowledge Network Systems let users find experts in the company that have already solved a problem. These can be a simple as letting users post a question to a discussion forum.
25
Knowledge Work Systems
* Knowledge workers include researchers, designers, architects and engineers whose main function is to create new knowledge in the form of ideas.
* Knowledge workers rely on office systems but also require very specialized Knowledge Work Systems.
26
Characteristics of many Knowledge Work Systems.
– Fast processing
– Sophisticated graphics
– Analytical tools
– User friendly interface
– Ability to collaborate with other Knowledge Workers.
27
Types of Knowledge Work systems.
– A plant engineer might use a 3D CAD package.
– A petroleum engineer may use a geophysical interpretation software.
– A marketing analyst might use a statistical analysis and inference package.
28
Intelligent Techniques
* Intelligent techniques rely on Artificial Intelligence and database technology to find relationships and patterns in mountains of data. This is called Knowledge discovery.
* Intelligent techniques also try to capture tacit knowledge resident in experts and try to reproduce it in rule based expert systems
29
* Some Types of Intelligent Techniques
– Expert systems
– Case Based Reasoning
– Artificial Intelligence
* Fuzzy Logic
* Neural Networks
* Genetic Algorithms
– Intelligent Agents
30
Types of Intelligent Techniques Expert Systems
* Systems used to capture the decision making logic of an acknowledged expert.
* This logic is often expressed through the use of IF, THEN, ELSE Rules.
* Hundreds of these rules are combined into a Knowledge base.
* The Expert System then “asks” a “novice” a series of questions that branch through the rules to an appropriate answer/recommendation.
31
Types of Intelligent Techniques Expert Systems pt 2
* Expert Systems are designed to work in narrowly defined problem domains.
* Granting a loan, diagnosing an illness, admitting a student, etc.
* However, as the environment surrounding the problem domain changes Expert systems have to be updated. Expert Systems do not “learn”
* For example and ES that makes ‘good” loan recommendations when interest rates are 2% will make “bad” recommendations when rates rise to 10%.
* A human expert would have adjusted his decision rules perhaps subconsciously. An Expert System can not do this. It has to be changed manually.
* Often the cost of maintenance exceeds the cost of development.
32
Types of Intelligent Techniques Case Based Reasoning
* Past experiences with situations, including the decisions and outcomes are captured and stored as cases.
* When a new situation is encountered, cases that closely match the new situation are retrieved.
* IF the decisions in the old case led to success AND if the new case is very similar to the old THEN the same decisions are made. ELSE a new decision is made.
* The new case, along with the decisions, outcomes and perhaps commentary is added to the case repository.
33
Types of Intelligent Techniques Artificial Intelligence
* AI is a branch of computer and cognitive sciences that look at how people learn with the goal of building “learning machines”
* AI uses many techniques to create these artificial learning machines. These include: – Fuzzy Logic– Neural Networks– Genetic Algorithms
34
Types of Intelligent Techniques AI – Fuzzy Logic
* Fuzzy Logic is “a rule based technology that can represent imprecision”.
* Fuzzy Logic uses advanced set theory to define concepts like HOT, COOL and COMORTABLE so that they can be interpreted by a digital computer.
* Fuzzy Logic is popular in Europe and Japan where it is used in cameras, antilock breaking systems, cooling systems, etc.
* Laudon also notes that it is used in Wall Street to identify acquisition targets.* Traditional Computer says IF TEMP >=90 THEN TURN FAN ON * Fuzzy Logic says IF TEMP = HOT THEN TURN FAN ON= COMFORTABLE.
35
Types of Intelligent Techniques AI – Neural Networks
* Neural Networks emulate the Neurons, Axons, Synapses and Dendrites found in human brain cells.
* We receive input from our senses that cause these Neurons to send electrical impulses to other Neurons.
* Eventually as the stimulus is repeated “hard wired” paths form between some Neurons.
* The brain “learns” to activate these “circuits” depending on the stimulus.
36
Types of Intelligent Techniques AI – Neural Networks pt 2
* Neural Net researchers hope to mimic the brain by teaching artificial neurons, etc. to fire in particular ways when they receive an input.
* The trick in Neural Networks is to teach these circuits to compare new stimuli to old “leaned” stimuli and fire/not fire depending on how close the old matches the new.
* This is how “learning” takes place.
37
Types of Intelligent Techniques AI – Neural Networks pt 3
* Neural Nets are particularly suited to discovering patterns and relationships in data so complex that they can not be perceived by humans.
* So AI systems are often used in data mining applications alongside statistical applications.
38
Types of Intelligent Techniques AI – Genetic Algorithms
* Genetic Algorithm researchers also turn to the biology for inspiration. However, they seek to mimic the “learning” that takes place in evolution.
* With GA we start with a set of random solutions that are represented genes or chromosomes. We also have a “fitness test” that ranks each of these solutions based on “fitness for purpose”
39
Types of Intelligent Techniques AI – Genetic Algorithms
* The GA then discards the “least fit” and keeps the “most fit” It can then combine or mutate the “most fit” and again compare the new solutions with the fitness test.
* Thus the Genetic Algorithms, by trial and error learns what works and what does not work.
* In business we can theoretically construct fitness tests based on cost, profits or speed and represent possible solutions as genes. The system may then be able to select and combine “good” solutions.
40
Types of Intelligent Techniques Intelligent Agents
* Intelligent Agents are software programs that work in the background using a simple knowledge and rule base. The Agent will then take actions based on the pre specified rules.
* Book the cheapest direct flight between A and B. Buy the book from the store that offers the lowest combined price and shipping cost.
41
Management Challenges
* There are several challenges that companies may encounter when trying to deploy KM Systems.
– Insufficient resources and attention paid to structuring and updating the repositories.
– Too much poor quality content may end up in the repositories if screening mechanisms are not implemented.
– Lack of context makes it difficult to relate documents to each other.
– Perhaps up to 80% of what employees “know” is in the form of “Tacit” knowledge which is very difficult to capture and codify.
– Employees may not naturally share what the know. Structures and rewards must be put in place to encourage sharing.
– The search engines which extract content from the repositories can sometimes return too many irrelevant hits. This is a sign that insufficient care has been paid to structuring, pruning and tagging.
42
Management Solutions
* There seem to be some rules of thumb to consider when deploying a KM System.
– Deploy and validate in Proof of Concept Pilot projects.
– Choose a high value business process
– Test with the appropriate audience
– Measure ROI of Pilot and if possible extrapolate to enterprise wide ROI
43
Management Solutions
* While Piloting a KM System, managers should focus on the following metrics.
– Number of visits to the portal– Number of answers found in the knowledge base
– Number of answers provided by experts
– User ratings of answers on a 1 to 5 scale.
Management Information Systems
flashcards
Decks in class (9)
# Cards
