| Eureka Knowledge Sharing System: Case Study
 |
| The Eureka team drew
on a seminal PARC ethnographic study that
revealed how technicians used “war
stories” to teach each other to
diagnose and fix machines. |
PARC researchers developed the Eureka system in the mid-1990s to support a community of service technicians
in capturing, annotating, sharing, and updating
knowledge about fixing equipment in the field.
Eureka, which has significantly improved service
productivity and customer satisfaction and
lowered service costs, is currently in daily
use by about 20,000 Xerox technicians worldwide.
In contrast to conventional
training systems that emphasize top-down delivery
of information, Eureka leverages social practices
for building knowledge. The system promotes
knowledge sharing via “tips,”
submitted by service technicians, which identify
machine problems and propose solutions. The
tips comprise a knowledge base that can be
accessed and enhanced by all members of the
community. Instead of financial incentives,
Eureka relies on recognition to motivate workers
to share knowledge and maintain the tips knowledge
base, which currently contains nearly 50,000
tips.
 |
Begun as a collaboration
between PARC researchers and Xerox France,
Eureka and its enabling technologies were
developed through several iterations using
a socio-technical methodology. Originally
conceived as an artificial intelligence (AI)
expert system that would guide users toward
the “best” solutions to machine
problems, Eureka’s initial developers
were a team of engineers, mathematicians,
and computer scientists. Technicians didn’t
find the smart diagnostic tool useful because
it offered expertise only on standard solutions
to common problems. What technicians needed,
instead, was a way to diagnose difficult and
unusual problems.
War
Stories
The AI team drew on
the findings of a seminal PARC ethnographic
study that revealed how technicians used “war
stories” to
teach each other
to diagnose and
fix machines. The
researchers altered
their course, turning
AI on its head
by turning
the work community
itself into Eureka’s
“expert system.”
Eureka’s effectiveness
is not based on the sophistication of its
technology but on its insights about how and
why people share knowledge. The system provides
a technological infrastructure as well as
processes and practices that enhance lateral
communication among technicians. It uses a
knowledge base consisting of cases in a simple
problem-cause-solution structure, with links
to other documents and pictures, and a very
rapid search process for locating relevant
tips. It also includes processes for validating
new tips and an infrastructure for distributing
them so that they can be used when technicians
are off-line.
Improved
Productivity
PARC researchers worked
closely with service technicians to design
the system, which was originally deployed
on France’s Minitel national communications
infrastructure. As a result of Eureka, France
moved from an average to a benchmark performer,
with service metrics that were better than
the European average by 5-20%, depending on
the product.
Eureka was scaled up and
transferred to a client-server system for
implementation in Canada and later in the
United States. By 2001, the number of problems
solved using Eureka had increased to nearly
200,000 annually. Since every solution can
save several hours of down time, possible
escalations to experts, and sometimes the
replacement of a machine, the use of Eureka
provides many millions of dollars in savings
annually for Xerox, and has led to increases
in both customer and employee satisfaction.
Research
Issues
In the Eureka design process, researchers
addressed a set of closely interrelated problems.
These had to do with community membership
and relationships, the kinds of knowledge
community members regularly shared with each
other and their motivations for doing so,
the existing practices and contexts through
which this sharing took place, and the process
for implementing a socio-technical system
to enhance these practices.
Eureka demonstrates the
value of participatory deployment and the
importance of making use of knowledge collected
on the front lines throughout an organization.
New
Eureka-Based
Systems
Eureka is the basis
for other community
knowledge-sharing
systems. One example
is LinkLite™,
which leverages
community-constructed
knowledge
bases in different
work settings.
The Eureka document collection
is also serving as a research corpus for PARC
scientists who are developing tools to help
automate the process of maintaining document
collections. They are applying these tools
to detect redundancy, contradiction, and obsolescence
in the Eureka tips collection.
|
