>
« back to results for ""
Below is a cache of http://www.worldinternetcenter.com/Think_Tanks/TTS2001/TTS01Nov13/11_13_01_SCM_Proceedings.pdf. It's a snapshot of the page taken as our search engine crawled the Web.
The web site itself may have changed. You can check the current page or check for previous versions at the Internet Archive. Yahoo! is not affiliated with the authors of this page or responsible for its content.
Artificial Intelligence (AI): Copyright ©2001 The Silicon Valley World Internet Center
November 13, 2001
P r o c e e d i n g s
The Silicon Valley World Internet Center and its programs are supported by:
Amdocs, Inc. Cable & Wireless Deutsche Telekom IBM Corporation
Fujitsu SAP Sun Microsystems
Artificial Intelligence (AI):
Is it Real?
AI & Supply Chain Management Copyright ©2001 The Silicon Valley World Internet Center
1
Executive Summary
On November 13, 2001, the Silicon Valley World Internet
Center convened a Think Tank Session, gathering partici-
pants, all active practitioners in the field, to discuss and
debate what the real opportunities are for Artificial
Intelligence (AI) driven implementations in Supply Chain
Management (SCM) systems. Given that the exact defini-
tion of AI is subject to ongoing debate, and in order to
provide a common ground for this session, AI was
defined as software that automates tasks that require
human intelligence, specifically, for the area of supply
chain management. SCM encompasses the organization
of, and supervision over the flow of materials, informa-
tion and finances as they move along the chain from the
supplier to the manufacturer and on to the distributor
and customer or end-user. SCM includes the coordina-
tion and integration of such flows both within the com-
pany and among business partners.
Think Tank participants discussed ongoing implementa-
tions of AI that are being leveraged to automate and
increase efficiencies in eCommerce. Based mostly on
personal, first-hand experience, participants agreed that
current efforts are focused more on demand, buy-side
types of applications rather than on supply-side imple-
mentations that require more intensive and costly back-
end integration. The proliferation of SCM systems has
created an abundance of data, with some experts going
as far as to say that there is more data available to indus-
try at this point in time than could be effectively used as
knowledge. Not much is being accomplished with the
data available, because, as one participant commented,
there is "too much data, not enough information." To
leverage available data, intelligent agents must be
deployed however, effective deployment depends on
the adoption of universally accepted standards.
Participants cautioned that standards are not a panacea.
Challenges to SCM were discussed at length.
Challenges were found to encompass four main areas
technology, organization, operations and finance.
Technological challenges include issues that pertain to
data, integration and standards, dynamics, complexity,
security and privacy, and the relative immaturity of artifi-
cial intelligence as a discipline. Organizational challenges
include change management, misalignment in objectives
and incentives across functions and divisions, and inter-
nal user resistance. Operational challenges stem from
the massive planning and scheduling effort required to
deploy a SCM system. Finally, economic and financial
concerns center around the need to establish clear ROI,
especially in the current recessionary environment.
Turning to predict the future, our panel of experts
considered the opportunities inherent in leveraging AI to
develop more efficient SCM systems and applications.
Within 12 to 18 months, participants predicted that
opportunities would revolve around data and process
integration, collaboration and integration tools, and facili-
tation of "human" activities. However, significantly more
opportunity was seen in the 18 to 36 month time frame,
including applications that would facilitate decision sup-
port, business rules, inventory management and the
leverage of data, improvements in man-machine inter-
faces, and further facilitation of human activities. The
real revolution would come five to ten years from now,
when semantic Web technologies will be employed to
dramatically increase supplier to manufacturer efficien-
cies.
Our panel concluded that while implementations of AI
in SCM are not about to replace humans, nor will they
likely do so in the foreseeable future, significant strides
have already been made in leveraging AI-driven technolo-
gies to increase efficiencies across the enterprise supply
chain. Future progress will occur incrementally as intelli-
gent agents proliferate and universal standards take hold,
enabling both up-stream and down-stream collaboration
and interoperability. Introduction
Over the past several years artificial intelligence (AI) has
become an inseparable part of both business and con-
sumer applications. We all now use AI-driven applica-
tions on a daily basis. But is AI being utilized to auto-
mate and increase efficiencies in eCommerce? Are there
valid models for the integration of AI in supply chain
management (SCM) today? And what will be the high-
est ROI (Return on Investment) SCM applications in the
short-term (12 to 18 months), medium-term (18 to 36
months), and long-term (36 to 60 months)? What oppor-
tunities would you pursue right now if you could? On
November 13, 2001, the Silicon Valley World Internet
Center convened a Think Tank Session to elucidate what
the real opportunities are for AI-driven implementations
in Supply Chain Management (SCM) systems. The center
gathered participants, all active practitioners in the field,
to offer, discuss and debate potential answers to the
above questions. The paper in front of you now summa-
rizes the proceedings of this session, providing a concise
view on how experts in the fields of AI and SCM see the
interface between these two important disciplines evolv-
ing over the next 12 to 60 months.
Think Tank Session participating companies included:
Coopetition, Inc., Dejima, Frictionless Commerce, Inc.,
Hewlett-Packard, iSpheres, Lake Forest Venture
Management, Open Run, Oracle Corporation, Radical
Data Designs, Rod Heisterberg Associates, SAP, SRI
International, Stanford University, Technology &
Strategy, United Parcel Service, VerticalNet, and Zesati.
Artificial Intelligence
Artificial intelligence is all around us yet many will tell
you it is nowhere to be found. An ongoing conflict rages
over coming to terms with a definition of artificial intelli-
gence that is acceptable to a majority of practitioners and
academics. The industry adage seems to be that if it is
practical, it cannot possibly be AI. However, AI now per-
meates such a variety of applications, both consumer-ori-
ented and of a business-to-business character, that it is
difficult to avoid. Artificial intelligence is embedded in
the voice recognition
engines that help us
navigate previously
incomprehensible and
frustrating menu trees
on customer service
calls. It is integrated
into the search engines
we use to navigate the
World Wide Web, as
well as into heuristic
search processes such
as at MapQuest a Web
service hardly associated with AI in ones mind. It is a
critical element in the success of Optical Character
Recognition (OCR) systems. And, in the context of
Supply Chain Management, AI has been put to extensive
use - for example, within the context of scheduling and
optimization applications at ILOG, for product configura-
tion at Trilogy, and in order to create speech interfaces
to enterprise data at Dejima.
What, then, is artificial intelligence? Clearly there is no
one set and well-accepted definition. Artificial intelli-
gence is commonly thought of as the simulation of
human intelligence processes by computer systems. An
artificially intelligent system should be able to learn and
reason. A common test for artificial intelligence is the
Turing Test, named for Alan Turing, a British computer
scientist. The Turing Test stipulates that in order for a
machine to be considered intelligent it must be able to
deceive a human being into believing that the machine is
Copyright ©2001 The Silicon Valley World Internet Center
2
A large factor in
todays economic
downturn has been
an excess of
manufacturing
inventory due to
unreasonable
long-term demand
forecasts. Inventory
is expensive -- 20%
to 40% of unit value
when annualized.
1
More information about Big Blue and the rivalry between Kasparov and the machine is available at
http://www.research.ibm.com/deepblue/home/html/b.html. human. Aside from Gary Kasparovs remarks about Big
Blue being intelligent following his narrow 1997 loss to
the machine