Chapter 7: Online Matchmaker eHarmony (57 questions)

glassesbeepingAI and Robotics

Oct 20, 2013 (3 years and 9 months ago)

278 views

Chapter 7:

Online Matchmaker eHarmony (57 questions)

Chances are you haven't heard of Dr. Neil Clark Warren, but you'd probably recognize him if you
saw him. He is one of America
's best
-
known experts on singles' issues, mate selection, and
developing healthy relationships. He's written 9 books, published hundreds of articles, and appeared
on more than 4,000 radio and television programs. Even with all these credentials you probabl
y still
don't know who he is by his name alone. But if you watch television, you would know him as the
smiting guy on the eHarmony commercials who can't wait to help you meet your perfect mate
through his scientific patented Compatibility Matching System.

The eHarmony Compatibility Matching System is an expert system that has been programmed with
Dr. Warren's knowledge about what makes a good relationship. Its goal is to match partners for
successful relationships using 29 key dimensions that help predict
compatibility and the potential
for relationship success. The 29 key dimensions are organized into two general categories: Core
Traits that include emotional temperament, social style, cognitive mode, and physicality, and Vital
Attributes that includes re
lationship skills, values
,

and beliefs, and key experiences.

An eHarmony applicant fills out a 436
-
question Relationship Questionnaire that allows the expert
system to categorize that person's personality and build a compatibility profile. Heuristics are
used
to search the eHarmony database for compatible partners whose personality attributes make a good
match according to Dr. Warren's studies. An ordered List of potential partners is produced with the
best candidates listed first. The service then provide
s methods to get in touch with prospective
mates.

Is the system successful? eHarmony has more than 11 million registered users and, according to an
independent poll, more than 90 eHarmony members on average marry every day as a result of being
matched th
rough the service. eHarmony uses artificial intelligence throughout its organization.
Technology c
u
tted predictive analytics, from SPSS lnc
.
, is used in various areas of the eHarmony
business including scientific research, brand development, product resea
rch, compatibility models,
customer satisfaction and retention, and projective analysis. Predictive analytics is a form of data
mining that employs artificial intelligence techniques to make assumptions about the based on
historical data and to predict out
comes of events. You could imagine how such technology could be
applied to matchmaking as well as traditional business activities.

eHarmony’s Senior Director of Research and Product Development Steve Carter is a strong believer
in Al for business, He belie
ves that predictive analytics is the lens through which eHarmony views
at of its data. The numerous analytic and data management tools provided by the SPSS systems
enables eHarmony to understand important information in more novel, forward
-
thinking ways.

eHarmony is a prime example of how artificial intelligence and expert system tools are taking a Lot
of the guesswork out of Life.


Discussion Questions

1. What role does artificial intelligence play in the matchmaking process at eHarmony?

2. How does eHar
mony make use of predictive analytics in all of its business units?




Critical Thinking Questions

1. Do you think the scientific methods provided by eHarmony are superior or inferior to traditional
random chance encounters for finding a mate? Why?

2. What

dangers to privacy and safety, if any, are involved in using a service like eHarmony?


SOURCES: DM Review Editorial Staff, 'eHarmony Expands SPSS Deployment Company
-
Wide
for Research and Development, ' DM Direct Newsletter, June 9, 2006, www
.
dmreview.
com/article
_
sub.cfm?articte
l

D= I 055723. eHarmony Web site, accessed June 30, 2006,
w.eharmony.com.



Call Centers Use Artificial Intelligence to Improve Service to Customers

Call centers, sometimes referred to as contact centers are emerging as an import
ant tactical weapon
for corporations in the battle for the customer dollar. A call center is a unit within an organization or
an outside firm that handles remote customer communications. Most 800 numbers you dial for
placing orders and after
-
sales support
are handled by a call center. Telephone solicitation is also
handled by call centers.

Internet
-
based technologies, such as e
-
mail, discussion boards, and chat, are increasing the duties
and boundaries of call centers. Research shows that about 70 percent
of real
-
time customer
interactions are handled by ca[[ centers. Business executives are picking up on these statistics and
investing heavily in new technology to empower call center operators and provide convenience to
customers. Artificial intelligence te
chnology plays a key role.

''Contact centers are entering a new phase and becoming more intelligent," analyst Catriona Wallace
says. ''New technologies will allow them to take an even more central role in organizations.
"Wallace, whose research company Cal
lcentres.net tracks growth in the sector, says increasing use
of the Internet by consumers, and technologies such as VOIP and speech recognition are driving
change.

Increasingly, the software used by call centers is being integrated with powerful CRM syste
ms. This
integration provides a solid foundation on which a range of new applications and agent techniques
can be built. With ready access to everything from a customer's purchasing history to personal
details and demographics, selling can take on a new di
mension.

For example, the outbound calling function of call centers has been hampered by consumer
resistance to unsolicited call and the rise of do
-
not
-
call lists. Companies need to be smarter about
when and why they contact the public. New tools enable co
mpanies to optimize the lists they are
using, and customize them for specific campaigns.
P
revious histories and details can be
automatically checked to increase the chance of finding a receptive customer, rather than randomly
dialing people.

Intelligent s
oftware too[s can examine demographic profiles to decide the best time to reach
customers and whether to call their work, home, or mobile number. Users of the software claim a 40
to 50 percent improvement in speaking to the right person at the right time a
nd the right place.

Intelligent software is also being used to accurately match the number of outgoing calls being made
to the number of agents in a call center. New versions of automatic dialers can predict how often
agents will become available and have
a caller waiting on the line as soon as they are free,
increasing agent utilization.

New intelligent software also helps companies respond smarter when customers or prospective
customers call them.

James Brooks, senior vice president at call center special
ist Genesys, says Leading companies are
examining a technique called psychographic routing. This involves routing incoming to the agent
that the system deems is the best match for the caller. Based on caller line
I
D, a psychographic
system can instantly as
sess known details such as the age, gender, and previous history of a caller,
and automatically route them to the appropriate agent. For example, a 55
-
year
-
otd shopping for
Life insurance would be routed to an operator of a similar age who is more to unde
rstand the
concerns of the caller.

Such systems can also provide important customer information relevant to targeted sales. For
example, the software can alert an agent that a customer is qualified for an increase in their credit
limit, which can then be o
ffered during the call. The success of such targeted sales pitches is high.

More complicated information could be calculated and provided. For example, bank call center
agents can be provided with real time credit scoring for customers seeking loans. "You
can also
undertake things Like predictive claims," says Tim Macdermid, Australia manager for analytical
software specialist SPSS. "An insurance company call center agent can collect information from a
customer and the system will predict whether that claim

is fraudulent."

From the customer's perspective, this call center evolution will lead to big changes in the way they
interact with large organizations, increasing amounts of transactions and business communications
will be conducted through call centers.
Call center agents will be more empowered as their toots
become more intelligent and customers will have better experiences when they dial 800 numbers.


Discussion Questions


1. How is intelligent software helping call center agents to be more effective?

2
. Why are call centers becoming the primary conduit for communication between a business and
its customers?

Critical Thinking Questions

1. What is psychographic routing and how might some interpret it as an infringement on their
privacy, whereas others
consider it a valuable service?

2. List communications technologies that a customer might use to communicate with a call center
agent along with the pros and cons of each method. How might intelligent systems assist with the
cons?


S0URCES: Grayson, lan, "
Digging Deeper into Data," Australian lT, April25, 2006,
http://australianit. news.com.aa. Genesys Corporation Web site, accessed June 30, 2006, www.
genesystab. com. SpSSWeb site, accessed June 30, 2006, www.spss. com.


SUMMARY

Principle

Knowledge managem
ent allows organizations to share knowledge and experience among their
managers and employees.

Know
l
edge is an awarenes
s and understanding of a set of
information and the ways that
information
can be made useful t
o support a specific task o
r reach a decision. A knowledge
management system
IKMS] is
an organized collection of peo
p
l
e, procedures, software
, databases and devices used to
create, store,
share, and use the organization
s know
ledge
and experience. Exp
l
icit know
ledge is
objective and
can be
measured and documented

in reports, papers, and rules. Tacit knowledge is
hard to measure and d
ocum
ent and is
typical
l
y not objective or forma
lized.
Knowledge workers are
peop
le who create, use, and
disseminate knowledge. They are usua
ll
y profession
a
ls in
science,
engineering, bus
i
n
ess, and other areas. The Chief
Knowledge Officer
(
CKOJ
)

is a top

l
eve
l

e
xecutive who helps
the organization use a KMS t
o create, store, and use knowl
edge to achieve
organizationa
l
goal
s. Some organizations and
professions

use communities of practice
(
CO
P) to
create,
store, and share knowledge. A C
OP

is a group of peop
le or
community dedicated to a
common discip
line or practice,
such as open
-
source software, auditing, medicine, eng
ineering, and
other areas. O
btaining, stori
ng, sharing, and using know
ledge is the
key to any KMS. The use of a
KMS often
l
eads to additiona
l
knowledge creation, storage,

sharing, and usage. Many tools
and
techniques can be used t
o create, store, aid use knowl
edge. These too[s and techniques are av
ai
l
ab
le
f
rom
IBM,
Microsoft, and other companies and organizations.


Principle

Artificial intelligence systems form a broad and diverse set of systems that can replicate
human decision making for certain types of well
-
defined problems.

The term artificia
l

intelligenc
e used to describe computers
with the abi
l
ity to mimic or dup
l
icate the
functions

of the
human brain. The objective of bui
l
ding
AI systems is not to
replace human decision
making
completely

but to rep
licate it
for certain types of we
ll
-
defined
prob
l
ems.

I
nte
ll
igent

behavior

encompasses

severa
l characteristics,
inc
l
uding the abilities to
l
earn from
experience an
d
app
ly this
knowledge to new experiences; hand
le complex situations
and so
l
ve
prob
l
ems for which

pieces of information might be
missing;

determine re
l
evant information in
gi
ven situation,
think in a
l
ogica
l

and rationa
l

manner,

and give a quick and
correct response; and
understand visua
l

im
ages and process
symbo
l
s. Computers are better than peop
le at transferring
information, making a ser
ies of ca
l
cu
l
ations rapid
ly and
accurate
l
y, and making comp
l
ex
ca
lculations, but human
beings are better than computers at
all other attributes of
inte
ll
igence.

Artificia
l

inte
ll
igence is a broad fie
l
d that inc
l
udes several

key components, such as e
xpert s
ystems,
robotics, vision
systems, natura
l

l
anguage processing, learning systems, an
d
neura
l

networks. An
expert
system consists of the hardware
and software used to p
roduce systems that behave as a
human expert wou
l
d in a specia
l
ized fie
l
d or area
(e.g., c
redit
analysis
)
. Robotics uses mechanica
l
or computer devices to
perform tasks that require a
high degree of precision or are
tedious or
hazardous for humans [e.
g., stacking cartons on a
pa
ll
et).
V
ision systems include hard
ware and
software that
permit
computers to capture
, store, and manipulate images
and pictures
(
e.g., face
-
recognition software). Natura
l lan
guage processing al
l
ows

the computer to understand and
react to
statements and commands made in a natura
l lan
guage, such as Eng
l
ish. Lea
rning syst
ems use a
combination
of software and hardware to a
ll
ow a computer to chan
ge how
it functions or reacts to
situations based on feedback it
receives (e.g., a computerized chess game
)
. A neura
l network i
s a
computer system that ca
n simulate the functioning o
f a
human brain (e.g., disease

diagnostics
system). A genetic
a
l
gorithm is an approach to so
l
ving Large, comp
lex problems
in which a number
of re
l
ate
d operations or models change
and evo
l
ve unti
l

the best one

emerges. The approach is
based
on the theory of

evo
l
ution, wh
ich requires variation and nat
ural

selection.

I
nte
ll
igent a
gents
consist of programs and a knowl
ed
ge

base used to perfor
m a specific task for a person,
a process, or
another program.


Principle

Expert systems can enable a novice to perform at the level of an expert but must be developed
and maintained very carefully.

An expert system consists of

a collection of integrated and
re
l
ated components, inc
l
uding a
knowledge bas
e, an infer
ence engine, a
n explanation faci
l
ity, a know
ledge acquisition
faci
l
ity, and a
user
in
terface. The know
ledge base is an exten
sion of a database, discussed
in Chapter3, and an
information
and decision support system,
discussed in Chapter 6. It con
tains a
ll

the re
l
evant da
ta,
ru
l
es, and re
lationships used in the
expert system. The ru
l
es are often composed

of if
-
then
state
ments, which are used for d
rawing conclusions. Fuzzy logic
a
ll
ows expert systems to
i
ncorporate facts and re
lationships
into expert system know
l
edg
e bases
that might be imprecise
or
unknown.


The inference engine proce
sses the rules, data, and rela
tionships stored in the kno
wledge base to
provide answers,
predictions, and suggestions the way a human expert wou
ld.
Two common
methods for
processing include backward and
forward chaining. Backward chaining starts with a
conc
lu
sion, then searches for facts
to support it; forward chaining
starts with a fact, then searches
for a conclusion to support it.

The exp
l
anation faci
l
ity of an expert s
ystem a
llows the
user to understand what ru|.es w
ere used in
arriving at a deci
sion. The knowledge acquisition faci
l
ity he
lps the user add or
update knowledge in
the kno
wledge base. The user interface
makes it easier to deve
lop and use the expert system.

The peop
l
e invo
l
ved in the deve
lopment of an expert sys
tem inc
l
ude the domain expe
rt, the
knowledge engineer, and the knowledge users
. The

domain expert is the person or
group who has
the expertise
or knowledge being captured for
the system. The knowledge
engineer is the developer
w
hose
job is to extract the expert
ise from the domain expert. The
knowledge user is the person who
benefits from
the use of
the deve
loped system.

The steps invo
l
ved in the deve
lopment of an expert system
inc
l
ude determining requi
rements,
identifying experts
, con
structing expert system components, imp
l
ementing resu
lts,
and mainta
ining
and reviewing the system.

Expert systems can be imp
l
emented in severa
l ways. Pre
vious
l
y, traditiona
l

high
-
l
eve
l

languages,
inc
l
uding Pasca
l,
F0RTRAN, and C0B0L, were used. L
ISP and PROLOG are two l
anguages
specifica
ll
y deve
l
o
ped for creating expert systems
from scratch. A faster and l
ess
-
expensive way io
acquire an
expert system is to purchase a
n expert system shell or exist
ing package. The she
ll

program is a co
ll
ection of software

packages and too
l
s used to design, deve
l
op, implement, and

maintain expert systems.


The benefits of using an expert system go beyond the typica
l

reasons for using a co
mputerized
processing solution.
Expert systems di
sp
l
ay "inte
lligent" behavior
manipu
late
symbo
l
ic information
and draw

conclusions, provide portable
know
l
edge, and can dea
l
with
uncertainty. Expert systems
can
be used to solve prob
l
ems in many fie
l
ds or discip
lines and
can assist in a
ll

stages of the
pro
b
lem
-
solving process. Past
successes have shown that expert sys
tems are good at strate
gic
g
oa
l

setting, p
l
anning, design, decision making, qua
lity
control and monitoring, and diagnosis.

App
l
ications of expert systems and artificia
l

inte
lligence
inc
l
ude cre
dit granting and Loan ana
lysis,
catching cheats
and terrorists, budgeting
, games, information management
and retrieval, Al and
exper
t systems embedded in products,
p
l
ant
l
ayout, hospitals and medica
l
faci
l
ities, he
lp desks and
assistance, emp
l
oyee
performance eva
l
uatio
n, virus detec
tion, repair and maint
enance, shipping,
and warehouse
optimization.


Principle


Virtual reality systems can reshape the interface between people and information technology
by offering new ways to communicate information,
processes, and express ideas creatively.


A virtua
l

rea
l
ity syst
em enables one or more users to move
and reac
t

in a compute
r
-
simulated
environment. Virtual reali
ty simu
l
ations requir
e special interface devices that
the sights, sounds,
and sensations of the

simulated
the user. These devices can a
lso record and send the speech
and
movements of the p
articipants to the simulation programs.
Thus, users are ab
l
e to sense and
manipu
late virtual
much as they wou
l
d rea
l

objects. This natura
l style of
action gives
the
participants the fee
ling that they are
immersed in the simu
l
ated wor
l
d.

Virtua
l

rea
l
ity can a
l
so refer to app
lications that are
fu
ll
y immersive, such as mouse
-
contro
lled
navigation
through a three
-
dimensiona
l
environment on
a graphics
monitor, stereo v
iewin
g from the
monitor via
stereo projection systems, an
d others. Some virtual reality
app
l
ications a
ll
ow views of

real environments with superim
posed virtua
l

objects, Virtua
l

rea
l
ity app
lications are found
medicine, education and tr
aining, real estate an
d tourism,
entertainment.


Principle


Specialized systems can help organizations and individuals achieve their goals.


A number of specialized sy
stems have recently appeared to
assist organizations and individua
ls in
new and exciting ways.
Segway, for
examp
l
e, is an e
l
e
ctric scooter that uses sophis
ticated software,
sensors, and gyro mo
tors to transport
people through warehouses, off ices, downtown sidewa
lks,
and other spaces.0riginatly desig
ned to transport people
around a factory or around town, more
recent versions are

being tested by the mi
l
itary for gathering inte
ll
igence and

transporting wounded
soldiers to safety. Radio
-
Frequency

i
dentification
(
RFID) tags are used in a variety of settings.

Game theory invo
l
ves th
e use of information systems to
deve
l
op competitive strate
gies for people,
organizations,
and even countries. Informatics combines traditiona
l disci
p
l
ines, such as science and
m
edicine, with computer science.
Bioinformatics and medica
l informatics are examples.
There are
a
l
so a number
of

special purpose telecommuni
cations systems that can b
e placed in products for
varied uses.



REVIEW QUESTIONS

1. Define the term artificial intelligence.

2. What is the difference be
tween knowledge and
information?

3.
W
hat is a vision system? Discuss
two applications of such
a system.

4.
W
hat is natural language processing?'

What are the three
levels of voice recognition?

5. Describe three examples of

the use of robotics. How can a
micro

robot be used?

6.
W
hat is a learning system? G
ive a practical exa
mple of such
a system.

7.
W
hat is a neural network? Describe two

applications of
neural ne
t
works.

8.
W
hat is an expert system shell?

9. Under what conditions is the deve
lopment
system likely to be worth the effort?

10. Identify th
e basic components of an e
xpert
describe the role of each.

1
1.
W
hat is fuzzy logic?

12.

W
hat

is

virtual

reality?

Give

several

examples

of

its

use.

13. Expert systems can be built based on rules or cases.
What
is the difference between the two?

14. Describe the roles of t
he domain expert, the knowledge
engineer, and the knowledge user in
expert systems.

15.

W
hat are the primary b
enefits derived from the use of
expert systems?

1
6.

What is informatics? Give a few examples.

17. Describe three applications
of expert systems o
r artificial
intelligence.

18. Identify three special interface devices developed for use

with virtual reali
t
y systems.

19. Identi
fy
, and briefly describe

three specific virtual reality
applications.

20. Give three examples of other specialized systems.


DISCUSSION OUESTIONS


1.

What are the requirements for a computer to exhibit

human
-
level intelligence? How long
will it be before we

have the technology to design such computers? Do you

think we should
push to t
ry to accelerate such a develop
ment?

W
hy or why
not?

2.

You work for an insurance company as an entry
-
level manager. The company contains both
explicit and tacit knowledge. Describe the types of explicit and tacit knowledge that might
exist in your insurance company. How you would capture each type of know
ledge?

3.

W
hat are some of the tasks at which robots excel?
Wh
ich

human tasks are difficult for them
to master?
Wh
at fields

of

AI are required to develop a truly perceptive robot?

4.

Describe how natural language processing could be used in

a universi
t
y setting
.

5.

Discuss how learning systems can be used in a military war

simulation to train future
officers and field commanders'

6.

You have been hired to dev
elop an expert system for a uni
versity career placement center.
Develop five rules a student

could use in selec
ting a career.

7. What is the purpose of a knowledge base? How is one developed?

8. What is the relationshi
p

be
t
ween a database and a knowledge base?

9. Imagine that you are developing the rules for an expert

system to select the strongest
candidates for m
edical school.

What rules or heuristics would you include?

10. Describe how informatics can be used in a business

11. Which interface is last developed and most challenging to create virtual reality system?

W
h
y
do you think this is so?

12. What application

of virtual rea.liry has the most potential

to genera
t
e increased profits in the
future?



CAREER EXERCISES


1.

Using the Internet or a library, explore how expert sys
t
ems

can be used in a business career.
How can expert sys
t
ems

be used in a nonprofit company
?

2.

Describe the future of artificial intelligence in a career are of your choise.


CASE STUDIES


Case One

BMW Drives Virtual Prototypes


Today's cars consist of about 20
,000 parts, and through virtual
rea
l
ity, BMW assemb
l
es the parts

and examines how they
interact in a variety of simu
l
ations in virtua
l space. The anal
ysis takes place
in BMW's vi
rtual reality studio, nicknamed
the Cave, wh
i
ch features a 175
-
square
-
foot PowerWa
ll
dis
p
l
ay and some serious computing power.

Among the many uses of t
he VR system,
the most valuable is simul
ating test crashes. After an
automobi
le is designed,
and prior to bui
l
ding a pro
totype, technicians wearing 3
-
D
glasses can
observe how

a crash affects a vehic
le down to the
smal
l
est detai
l
s. With this in
formation, they
recommend
changes to the design to improve driver and passenger safety.

Before a new mode
l

is assemb
led, it has been crashed in
a hundred different

ways in
BMW's Cave.
Smashing the virtual
car into a virtua
l

wal
l
t
a
kes two to four days of
computing time.
The
computer
works arou
nd the clock, breaking down the
moment of impact into stages of mi
ll
iseconds. The
resu
lting
3
-
D fi
l
m is viewed by technicians in super

stow motion. At any
point in the crash, the
vehic
le can be dissected to examine
the effect of the cras
h on particu
l
ar automobi
le areas and
components.

Prior to virtua
l

reality,
BMW tested actual prototypes of
new mode
l
s. Such prototypes c
an cost up
to a million dollars
to manufacture. Each virtua
l

3
-
D simu
lation costs under
$500. After a vehic
l
e's
desig
n h
as been improved based on VR
crash tests, actua
l

prototypes are manufactured and a
lso
crashed. BMW must prove in rea
l Life the theories derived
from the virtua
l

rea
l
ity simu
l
ations.
Computer simu
lations
are on
l
y valid to the extent that they ref
l
ect rea
lit
y. At one time,
BMW might
have test
-
cras
hed six prototypes. Today, they
crash on
l
y a coup
l
e and have the additiona
l

benefit of
resu
lts
from the study of over 100 VR crashes.
In this way, they can
produce safer vehic
l
es for less
money.

Safety is on
l
y one of

s
everal uses of BMW's VR system.
Designers use the system to experiment
with alternative interior and exterior designs. Surfaces of meta
l
, artificia
l mate
ria
l

and Leather are
rea
l
istical
l
y represented in virtua
l space
to have rea
l

t
extures and

properties
such as reflection.
Designers can examine a vehic
l
e

inside and out in a variety of l
ighting scenarios, changing colors,
and materia
ls as desired
to find t
he perfect combinations.

Another department te
sts add
-
ons and pays particular
attention to ease of ope
ration. For examp
le,
how much energy
does it take to close the top

of a convertible? What happens
when you s
l
am the
trunk Lid? A technician can repeated
l
y

c
l
ose a virtua
l

trunk Lid, using
varying degrees of force,
carefull
y analyzing how the meta
l

reacts e
ach time. How difficu
lt
is it to open a door? Recent
ly,
BMW engineers tried out the
door of a new virtua
l

mode
l and heard a subdued, high
pitched
meta
ll
ic ping when
the door was shut. They thought
that the sound seemed cheap and tinny.
Rea
lizing that a cus
tomer's opinion of a car is
often shaped when they open and
c
l
ose the door, they
set out to fi
nd the source of the ping. In a
matter of days, they had rede
signed the latch in the virtuaI
mode
l

to e
l
iminate the ping
. Before VR, such an alteration
would have

taken months of work at a
test bench.

BMW is a
l
so providing
training to its technicians and
mechanics using virtua
l

rea
l
ity. A new VR
training fac
ility for
BMW technicians in Unter
schleissheim, Germany, dramati
cally increases the
number of

technicians who

can be trained
at the same time.
I
nstead o
f each trainee working one at a
time, teams of technicians can now s
lip on high
-
tech goggles
and g
l
oves to work on a virtua
l

im
age
of a car or car part that's
projected onto the inside of the gogg
l
es.


Discussion
Questions

1. List

three

ways

in

which BMW

uses

VR in

the

design of new
mode
l
s of vehicles.

2. How does VR save BMW money over the long run?


Critica
l

Thinking Questions

l. How might VR be used in

your career area to save money
and provide opportunities not

avai
l
ab
l
e
in the rea
l
wor
l
d?

2
. When might the resu
l
ts of a

simulation in virtual space be
considered valid and usab
le without
running the same test
in rea
l

space?


SOURCES: Staff. "SG
I

Case Study: Virtua
l

Rea
lity: Crashes Without Dents,'
accessed Ju
l
y 1,
2006,
www.cgi.com. "BMW

Virtua
l

Rea
lity Room," accessed July 1, 2006,
www.ca
rpa
ges.co.uk/bmw/bmw_virtuat_reality_room_04
-
09_04
.asp. BMW Group Web site,
accessed July1, 2006,
www.bmwgroup.com
.


Case Two


Pacific Gas
and Ele
ctric Corporation Increases Cus
tomer Satisfaction with Speech Recognition

"
O
ur

vision is to become the Leading uti
lity in the United States
by de
l
ighting our 15 mil
l
ion
customers, energizing our

20,000
emp
l
oyees, and rewarding

our shareholders," wro
te Peter A.
Darbee, chairman of the b
oard, CE0 and president of Pacific
Gas and E
lectric (
PG&E
)

Co
rporation
in a recent letter to
company stakeho
l
ders. 0ne

way that PG&E is delighting it
customers is through
adding customized speech
-
recognit
ion
techno
l
ogy
to its existing inte
ractive voice response (IVR)
sys
tem in it
s customer service call center.

IVR is a computerized system that a
l
lows a ca
ller to inter
act with the system by sp
eaking
commands or other infor
mation. Typical
l
y used at the front end of ca
ll centers, an I
system might
ask you to en
ter or say" the extension of the
person you want to speak wit
h. lf you say "three nine
sever
two," the IVR system imp
lements speech
-
recognition tech
niques to
tr
ansform the w
ave
pattern of the sound of the
spoken n
umbers to actua
l

n
umbers that it uses as input. If
you are asked
a yes or no q
uestion, the system might be pro
grammed to accept 'yes," no," "yeah," "yep," 'nope,"
"oka
y”,
"no thanks," or even "ahuh," Some of the Latest systems u
se
natura
l

l
anguage processi
ng
techniques to a
llow callers t
ask questions without promp
ting for specific keywords. lf
system asks
the customer th
e purpose of the call., and the
ca
ll
er responds with "the
battery in my notebook
computer
longer ho
l
ds a charge," th
e system interprets the

statement
and rep
l
ies, "Ho
l
d the Line
please whi
le I connect you with notebook power specialist."

PG&E insta
ll
ed the cu
stomized speech
-
recognition sys
tem to he[p automate account

identification
and provide other
customer se
lf
-
service functiona
lity. The s
ystem was devel
oped with ScanSoft lnc.
and

Nortel Networks Corp. to run
a Windows server. The spe
ech
-
recognition system is inte
grated in
the company's Norte
l IVR system, along with its
customer information, out
age management, and
field order
schedu
l
ing
systems, said Steve Phil
lips, manager of PG&E
Corp's contact center
enhancement. The purpose of the
imp
l
ementation was to impro
ve customer satisfaction and tl
uti
l
ity's "Techno
l
ogy Take Rates.'' Techno
l
ogy Take Ra
tes
refers to the percentage
of customers
w
hose needs are sat
isf
i
ed using the IVR system

without speaking directly to a
customer service
representative.

Based on customer sur
veys, the speech
-
recognition sys
tem has improved PG&E's

customer
satisfaction and Tech
no
l
ogy Take Rates, sai
d Kent Barnes, a
senior project
manager in PG&E's
conta
ct center enhancement group. Over
the first year, the percentage
of customers rating the IVR
sys
tem as "exce
lle
nt" or "very g
ood" rose from 61 percent to
69 percent. The new speech
-
recognition techno
l
ogy enab
le
PG&E Co
rp to improve its techno
logy take rate from 33 per
cent to
about 38 percent.


Equa
ll
y important are the
savings that PG&E realized with
the new system. Based on an

average
cost increase of $5 to
$9 for a PG&E customer ser
vice representative to handle a
ca
ll

rather than the
IVR system, th
e utility's $3 million invest
ment in the speech software and associa
ted hardware was
paid off in less than a year.

I
nvestments in customer se
lf
-
service systems will remain
one of the most popu
l
ar are
as for
investment by
power com
pany
I
T operations, Gartner
I
nc. ana
lyst Zarko Sumic said in
a research
note. Uti
l
itie
s are eager" to reduce customer
support costs whi
l
e improving the qual
ity of service,
Sumic
said. "To meet higher comp
l
exity at the transformationa
l
leve
l

and in
crease R0l potentia
l,
leading energy companies
wil
l

comp
l
ement Web channe
ls with advanced voice tech
no
l
ogies."


Discussion Questions

1. What rote does speech rec
ognition play in modern
-
day IVR
systems?

2. How did the new spee
ch
-
recognition system save PG&E

Corp. money whi
l
e increasing customer
satisfaction?


Critical, Thinking Questions

1. Why do you think PG&E Corp's customers appreci
ated the
abi
l
ity to speak response
s rather than
keying them in or
wai
t
ing for an
op
erator?

2. As IVR systems become more
intel
ligent with broader
speech
-
recognition capabi
l
ities, how wi
ll
the call center
operator's duties change?


SOURCEST Hoffman, Thomas, "speech Recog
nition Powers Utitity's Custome Service,"
Computerworl
d,

S
eptember
12' 2005, www.computemorld com'
PG&E Cor
p Web site, accessed
Ju
l
y 1, 2006
, www.pgecorp'com Nortel Intell
igent Catt Manager Web site, accessed Ju
ly 1, 2006,
www.norteLcom
'
Microsoft Speech Server Web s
ite, accessed July 1, 2006,
www.microsoft.com


Questions for Web Case

See the Web site for this
book to read about the Whitmann
Price Co
nsulting case for this chapter.
Following are ques
tions concerning this Web case.


Whitm
ann Price Consulting: Knowledge
Management and Spe
cia
l
ized
I
nformation Systems


Discussion Questions

1. List three forms of Al th
at are being considered for the
AMCI system and how they are to be
used.

2. List the advantages an
d disadvantages of implementing
the Al systems in the AMCI s
y
stem.



Kulula.com: The Trials and Tribulations of a South African: Online Airline

Anesh Maniraj Singh

U
niversity of Durban

Ku
l
ula.com was
l
aunched in August 2001 as the first on
l
ine air
l
ine in South Africa. Ku
l
u
la
is one of
two airlines that are operated by
Comair Ltd. British Airways IBA), the other airlin
e
that Comair
runs, is a fu
ll
-
service franchise operation that serve
s the South African do
mestic market. Ku
l
u
l
a,
un
l
ike BA, is a
l
imited
-
service o
peration aimed at providing low
fares to a wider domestic m
arket
using five airc
raft. Since its inception, Kulula has rein
vented air trave
l in South Africa, making it
possibl
e for more peop
le to fl
y

than ever before
Ku
l
u
l
a is a true South African e
-
commerce success.
T
he company boasts as one of its
successes the
fact that it has been profitab
l
e from day one.
It is
recognized international
among the top low
-
cost airtines and participated in a co
nference attended
by other such
internationa
ll
y known
l
ow
-
cost carriers as Virgin B
l
ue, Ryanair, and easy Jet. Ku
l
u
la
also

received an award from the South African Department
of Trade and Industry for being
Techno
logy Top 1 00 company.

Ku
l
u
l
a's success is based on its c[earty defined s
trategy of being the lowest
-
cost
provider in the
South African domestic air travel industry
.

To this end, Kulula has adopted a no
-
f
ri
ll
s approach.
Staff and cabin crew wear simp
l
e uniforms, and the company h
as n
airport lounges. There are no
business c
l
ass seats and no frequent
-
f
lyer programs. Cus
tomers pay for their food and drinks. ln
addition
, Ku
l
u
l
a d
oes not issue paper tickets, an
very

few trave
l

agents book its ftights
-
90 percent of
tickets are so
l
d direct
l
y

to customers
.
Furthermore, customers have to pay for ticket changes, and
the company has a po
l
icy
of “
no f
l
y, no refund." Yet, in its
drive to keep costs down, Ku
l
u
l
a does
not comp
romise on
maintenance and safety, and it emp
l
oys the best pitots an
d meets the highest
safety stan
dards. Like a
ll

B
2C companies, Ku
l
u
l
a aims to create customer

va
lue by reducing
overhead
costs, inc
l
uding salari
es, commissions, rent, and consumab
l
es such as paper

and paper
based documents. Furthermore, by cutting out the midd
leman such as travel agents
Ku
l
u
l
a is ab
l
e to
keep prices low and save custome
rs the time and inconvenience of
having to pick up tickets fro
m
travel agents.
I
nstead, customers contro
l

the en
tire shop
ping experience.

Ku
l
u
l
a was the so
l
e provider of
l
ow
-
cost f
l
ights in South Africa unti
l

early 2004, whe
n O
ne Time
l
aunched a no
-
fri
ll
s service to compete head
-
on with Ku
l
u
la. Due to the high
price
e
l
asticity of
demand within the industry, any lower
ing of price stimulates a high
demand for flights. The
increase in competition in the
low
-

price end of the market has seen Kul
u
l
a decrease fares by up to
20 percent whi
te increasing passengers by over
40
percent. There, however, has been no brand
switching
. Kulula has grown in the market at the expense of others.

Apart from its low
-
cost strategy, Kulu
l
a is successfu
l because of its strong B2C busi
ness mode
l
. As
previous
l
y mentioned, 90 percent of its
reve
nue is generated from direct
sa
l
es. However, Ku
l
u
l
a has
recent
l
y ventured into the
B2B market by collaborating with Computicket and a f
ew trave
l

agents,
who can log on to the Ku
l
u
l
a site f
rom their company
intranets. Kutu
l
a offers fares at substantia
l

redu
ctions to b
usinesses that use it regularly.
Furthermore, Ku
l
u
l
a bases its success on three simp
le
principles: Any decision take
must bring in additional

revenue, save on costs, and/or

enhance
customer service. Tech
no
l
ogy contributes substantia
ll
y to these
three princip
l
es.

I
n its first year,
Kulula

used a l
ocally devel
oped res
ervation system, which soon rat out of
functional
ity. The second
-
generation system was AirKiosk, which was deve
loped in
Boston for
Ku
l
u
l
a. The system change resu
l
ted in an improvement
of functiona
l
ity
to
passengers. For examp
l
e,
in 2003,
Kulula

ran a promotion

during which tickets were sold
at ridicu
l
ous
l
y low prices, and the
system was overwh
elmed. Furthermore, Kulula expe
rienced a system crash that tasted a day and a
ha
l
f, which sever
e
l
y hampered sa
l
es
and
customer service. As a result, year two saw a revamp in a
ll
technology: All the hardware
was rep
l
aced, bandwidth was increased, new servers and database
servers were instal
led
and Web hosting was changed. ln short, the entire system
was reptaced.
Accor
ding to l'
Director Car
l

S
choltz, "
O
ur success depends on infrastructura
l

stabi
lity; our current
sys
tem has an output that is four times better than the best our systems cou
ld ever produce.
Ku
l
u
la
staff
members are conscious of the securi
ty needs of custo
mers and have invest
in 128
-
bit
encryption, giving customers peace of mind th
at their transactions and infor
mation are safe.

The success behind Kul
u
l
a's systems ties in its bra
nding
-
its strong identity in the
marketp
l
ace,
which inc
l
udes it
s name and visual appea
l
. The term
kulula means "easy,
and
Kulula
's

Web site has
been designed with a simp
l
e, no
-
fuss, user
-
friend
l
y interfac
e.
When visiting the Kulula si
t
e, you are
immediately awa
re that an airline ticket can be
purchased in three easy
steps. The first step a
ll
ows
cust
omers to choose destinations and
dates. The second step a
ll
ows customers to choose the mo
st
convenient or cheapest flight
based on their need.
Kulula

a
l
so allows customers to

book cars and
accommodations in
step two. Step t
hree is the transaction stage, which a
llows customers to choose
the most
suitab
l
e payment method. The confirmation and ticket
can be printed after payment have
been settled.
Kulula

has not embraced mobi
l
e comm
erce yet, because the technology
does not
suppo
rt the abi
l
ity to a
ll
ow customers to purcha
se a ticket in three easy steps. Unl
ike other e
-
commerce sites,
Kulula

is unc
l
uttered and s
i
mp
le to understand, enhanc
ing customer service.
Kulula is a f
un brand

-

with offbeat
advertising campaigns and bright
gre
en and b
l
ue corporate and
aircraft co
l
ors
-
but behin
d the fun exterior is a group
peop
l
e who are serious about business.

Ku
l
u
l
a's future is extreme
l
y promising. Techno
l
ogy changes continua
lly, and Kulula
strives to have
the best techno
l
ogy in p
l
ace at a
ll

times.

B
2B e
-
commerce wi
ll continue to
be a major focus of the
company to deve
lop additional
dis
tribution channels with little or
no cost.
I
n conjunction with bank
partners, Kulu
l
a is deve
l
oping additiona
l methods
payment to replace credit card payments,
a
ll
owing more peop
l
e t
he opportunity to fly.
These transactions wi
ll be f
ree. Ku
lul
a is a
l
so
invo
l
ving c
ustomers in its marketing effort
by obtaining their permission to promote specia
l
offers
by e
-
mai
l and short message se
vice to customers' ce
ll

phones.
The Kulu
l
a Web site wi
ll

soon serve
as a ticketing port;
where customers can a
l
so purchase British Airways t
ickets, in three easy steps.
The
company has many other developments in the pipeline that wi
ll enhance custom service.
According to Schol
tz, "We are

not an on
l
ine air
l
ine, just an e
-
tai
l
er

that setts air
l
i
ne tickets.
"


Discussion Questions


1. This case does not menti
on any backup systems, either el
ectronic or paper
-
based.

What would you recommend to ensure that the business runs

24
/
7
/
365?

2. It is c
l
ear from this case that Ku
lul
a is a Low
-
cost provider. What e
l
se cou
l
d Ku
l
u
la
with its
techno
l
ogy

to bring in additiona
l

r
evenue
, save on cost, and enhance customer
service ?

3.
Does the approach taken by Ku
l
u
l
a in terms of its strategy, its business mode
l
,
the three
princip
l
es of success
l
end itse
l
f to oth
er businesses wanting to engage
e
-
commerce?

4. Ku
l
u
l
a f
l
ights are almost a
l
ways fu
ll
. Do you think that by partnering w
ith a company
such as
Lastminute.com the airline cou
ld
f
l
y to capacity at a
ll times?
What are the risks
related to such a
co
ll
aboration?


Critica
l
Thinking Questions


1. Ku
l
u
l
a initia
l
ly deve
l
oped its systems in
-
house, which it
later outsourced to AirKiosk in
Boston.
Do you think it is wise f or an e
-
business to outs
ource its systems
development? I
s it strategica
ll
y
sound to outsource systems deve
l
opmen
t to a company in a different
country?

2. With the current trends in mobile commerce, cou
l
d Ku
l
u
l
a offer its services on mobi
le;
devices
such as ce
ll
u
l
ar phones? Wou
l
d the company have
to a
l
ter its strategic thinkin
g
to accommodate
such a shift?
I
s it possib
l
e to deve
l
op a t
ext
-
based interface that could
faci
l
itate a purchase in three
easy steps?