P1: IML/FFX P2: IML/FFX QC: IML/FFX T1: IML
WL040-Schalkoff WL040/Bidgolio-Vol I WL040-Sample.cls July 16, 2003 15:52 Char Count= 0
242 RULE-BASED ANDEXPERTSYSTEMSFigure 4: Strandware help session.SELECTED INTERNET APPLICATIONS
Web-Based Technical Support
Numerous examples of Internet-based technical support
exist. Software companies, in particular, observed early
on that many product technical support issues revolve
around problems that recur. Web-based technical sup-
port for these problems provides a number of benefits
for both the company and its customers. For example,
a certain percentage of customer support issues may be
solved autonomously. The customer gets a quick reso-
lution of the problem, and technical support resources
are conserved for the more challenging problems. In ad-
dition, autonomous Web-based technical support serves
as a filter for human technical support. When a cus-
tomer does not get a resolution to the problem using the
online expert system, human technical support may be
invoked. An additional benefit is that preliminary prob-
lem information has already been acquired by the au-
tonomous system and thus the duration required for
the session is shortened. Finally, the autonomous sys-
tem may gather all the data from many Web-based sup-
port sessions and thus provide a comprehensive database
of previously encountered problems and corresponding
solutions.
An example of this approach is Strandware, which
develops bar-code label design and data collection soft-
ware for industrial and business tracking applications
(http://www.strandware.com/). (A sample rule was shown
earlier.) The rule-based automated technical support sys-
tem is implemented using Amzi’s WebLS product. WebLS
is a simple ES tool, designed specifically for deploying ex-
pert systems on the Web.When the customer initiates a support session, he or
she is asked for information regarding the product and
problem. The system uses this initial information to filter
the rules that might apply and gathers additional infor-
mation as needed in its search for a problem resolution.
A sample screen is shown in Figure 4.
The Strandware ES is composed of questions, rules,
and answers. An initial goal starts the system searching
through the rules. The antecedents of some rules in this
system are facts with values that can be obtained by query-
ing the user. The following are two such facts related to
diagnosing a 16-bit ODBC (open database connectivity)
problem:question("second edition", [
prompt = $Is it Second
Edition or have you
installed the Y2K
update?<BR>$,
ask =yes_no,
askAfter = ["symptom"]
]).
question("BackTrack Version", [
prompt = $What version of
BackTrack are you
running?<BR>$,
ask = field,
askAfter = ['symptom']
]).Details of the Strandware system are provided at http://
http://www.amzi.com/customers/strandware.htm and in Hicks
(2000).