Expert Systems: Are they Right for your Business?
| DOI | https://doi.org/10.1108/eb057455 |
| Published date | 01 September 1986 |
| Pages | 18-20 |
| Date | 01 September 1986 |
| Author | Richard Heygate |
| Subject Matter | Economics,Information & knowledge management,Management science & operations |
Expert Systems:
Are they Right
for your Business?
by Richard Heygate
Director, Index International
Introduction
Expert systems are computer systems encoded with pieces
of human knowledge and experience. They make up
perhaps the most visible and promising branch of the nas-
cent Artificial Intelligence (AD technologies. Over the past
few years, the press has hailed artificial intelligence as it
emerged from academic laboratories and sought produc-
tive uses in the real world. Now such systems have cap-
tured the attention of IS and user managers; yet, few have
any direct experience with the technology.
Expert systems are used to solve problems by drawing on
human knowledge and experience. To do so, they perform
such activities as diagnosis, interpretation, planning and
design,
and are currently used in many fields, including
medicine, geology, finance and manufacturing.
Such systems have two primary components:
• a knowledge base encoded with the data, facts and
heuristics, or rule of thumb, about a particular sub-
ject area, and
• an inference engine containing the necessary
problem-solving procedures.
These components drive a typical session with an expert
system in the following way. A user is asked for informa-
tion needed to guide the search for a solution. As the user
provides the requested data, the inference engine combines
the information with the relevant components in the
knowledge base, discarding some possible solutions, pur-
suing others, and gathering additional information needed
to get closer to the solution.
Expert systems are built by specialists known as knowledge
engineers. Part systems analyst, part psychologist, a
knowledge engineer works with experts to draw out their
knowledge and insight and then encode it within a system.
Figure 1 indicates some key ways in which expert systems
differ from other technologies. They often run on specialis-
ed hardware and are written in unfamiliar programming
languages. So although the technology is promising, even
technically proficient managers have trouble evaluating it.
Figure 1. Comparison of Technologies
Technology 3rd Generation 4th Generation 5th Generation
Characteristics (MIS) (DSS) (ES)
Type of Numeric Numeric Symbolic
computation
Type of data Hard Hard Soft and hard
Delivery Batch On-line On-line
Cost Variable Low High
to medium
Development Variable Short Long
time frame to medium
Type of Structured Semi- Unstructured
problem structured
Who defines Analyst User Expert
the system?
Type of Data "What if" Consultation,
result scenarios explanation, or
solution
Technology Low Low High
risk to medium
As a result, many managers have fundamental concerns:
Where are the expert system opportunities? Are they worth
pursuing? How do I select the best opportunities? How do
I get started? Index system have developed a framework that
can help top management identify and exploit the best op-
portunities. Figure 2 summarises this three-step approach
for identifying, assessing and developing expert systems.
18 IMDS • SEPTEMBER/OCTOBER • 1986
To continue reading
Request your trial