This article was downloaded by: [UTSA Libraries]On: 07 October 2014, At: 07:37Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Organizational Computing andElectronic CommercePublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/hoce20

A Next Generation Multimedia CallCenter for Internet Commerce: IMCBong Ki Moon , Jae Kyu Lee & Kyoung Jun LeePublished online: 18 Nov 2009.

To cite this article: Bong Ki Moon , Jae Kyu Lee & Kyoung Jun Lee (2000) A Next GenerationMultimedia Call Center for Internet Commerce: IMC, Journal of Organizational Computing andElectronic Commerce, 10:4, 227-240, DOI: 10.1207/S15327744JOCE1004_02

To link to this article: http://dx.doi.org/10.1207/S15327744JOCE1004_02

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

A Next Generation Multimedia CallCenter for Internet Commerce: IMC

Bong Ki MoonJae Kyu Lee

Graduate School of ManagementKorea Advanced Institute of Science and Technology

Kyoung Jun LeeSchool of BusinessKorea University

Human assistance, as well as automated service, is necessary for providing more con-venient services to customers on the Internet-based commerce system. Call centershave been typically human-based service systems. However, the services of existingpublic switched telephone network-based call centers are not enough to meet theneeds of customers on the Internet. Most of them have been designed without consid-ering the interaction involved in shopping on the Internet. In our research, we design acall center named IMC (Internet-based multimedia call center) that can be integratedwith an Internet shopping mall. It contains 2 parts: an Internet multimedia dialoguesystem and a human agent assisting system. The Internet multimedia dialogue systemis an Internet and multimedia version of the interactive voice response service of com-puter telephony integration-based call centers because it provides access to the multi-media Web page along with the recorded voice explanation through the Internet. Thehuman agent assisting system aims to select the most appropriate human agents in thecall center and support them in providing high-quality individualized information foreach customer. IMC is a real-time, human-embedded system that can providehigh-quality services cost-effectively for Internet commerce.

call center, electronic commerce, Internet commerce,computer telephony integration, Internet shopping mall, human-embedded system

1. INTRODUCTION

Customers in the real world often want to talk with merchants over the telephone.To meet such needs effectively, companies have built customer service centers

JOURNAL OF ORGANIZATIONAL COMPUTINGAND ELECTRONIC COMMERCE 10(4), 227–240 (2000)

We appreciate the support of the Intelligent Information System Laboratory and the Internet Com-merce Laboratory at the Korea Advanced Institute of Science and Technology, LG–EDS Systems, andSambo Information Communication.

Correspondence and requests for reprints should be sent to Jae Kyu Lee, Graduate School of Manage-ment, Korea Advanced Institute of Science and Technology, 207–43 Cheongryang, Seoul 130–012, Korea.E-mail: jklee@msd.kaist.ac.kr

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

named call centers. A call center is defined as group of human agents and/or auto-mated voice response units that support communication with customers over thetelephone assisted by computers. The call center receives customers’ orders andclaims and answers customers’ inquiries and technical problems. Services of callcenters can be categorized into inbound service and outbound service. Inbound ser-vices respond to the customers’ calls, including taking orders, reserving, ticketing,taking care of after-service or claim requests, and providing information aboutproducts and services. Outbound service is an initiated activity by the call centers tocontact customers. Tasks include sales support, direct sales, area management,market surveys, public relations, advertisements, and so on.

Most current call centers use computer telephony integration (CTI) technology.CTI has made individualized and efficient telemarketing possible and has im-proved the productivity of human agents in the call center. In the past, the publicswitched telephone network (PSTN) has been used as the basic infrastructure ofthe CTI-based call centers.

As Internet technology evolves, the Internet infiltrates the PSTN’s traditionalarea of services such as telephone and facsimile. Many companies have developedInternet telephony products, so in the long run many services currently possibleonly on PSTN will be available on the Internet. We expect that without exceptionall call centers will make this change.

In this article, we deal with the call centers in the Internet commerce environ-ment. Human assistance, as well as automated service, is necessary for providingmore convenient services to customers on the Internet-based commerce systems.Call centers are typically human-based service systems. However, the services ofexisting CTI-based call centers are not enough to meet the needs of customers onthe Internet because most of them have been designed without considering the in-teraction involved in Internet commerce service systems such as in Internet shop-ping malls. In this article, we design a call center that can be well integrated with anInternet shopping mall.

The remaining sections are organized as follows: Section 2 explains the call cen-ters from various points of view. In Section 3, we discuss the necessity of the callcenter for Internet shopping malls through a scenario. In Section 4, we present theInternet-based multimedia call center (IMC) designed for Internet shopping mallcustomers. The prototype IMC–1 is demonstrated in Section 5. Section 6 concludesthis research and suggests further investigation.

2. CALL CENTER

2.1 Brief History of Call Centers

Call centers started their operations in early 1980s, and from a technical point ofview, they can be categorized into four generations.

1. The first generation of call centers provided information only or help throughthe phone. The legacy computer system and private branch exchange (PBX)worked separately. A human agent could retrieve a customer’s information fromthe legacy system only after the phone conversation had taken place. The human

228 MOON, LEE, LEE

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

agent had to provide all information to the customers through direct conversa-tion—a very friendly yet expensive option.

2. In the second generation, interactive voice system is and addition provided byintegrating the voice response units. Automating the human agent’s routine tasksalleviated his workload and made midnight ordering and booking service possi-ble. However, the human agent could only retrieve the customer’s informationmanually from the legacy system.

3. The third-generation call centers are the CTI-based call centers still usingPSTN. Integrating telephony technology with computer technology increased thecustomers’ satisfaction and the productivity of the call centers. In this generation,many additional efficient functions are supported such as intelligent call routingbased on the availability of human agents, automatic pop-up of customers’ infor-mation, simultaneous forwarding of both a call and a customer’s information to an-other human agent, predictive dialing in telemarketing, and so on.

4. The fourth generation is the Internet-based call center, which is the aim of thisstudy. As many parts of commerce move from physical markets to the Internet en-vironment, the infrastructure of the call center for parts should also change fromPSTN to the Internet because the Internet can provide more convenient and effi-cient services to the customers of Internet commerce. As the Internet shoppingmalls evolve, the role of the Internet-based call centers will become more impor-tant. Internet-based call centers can provide many convenient services such as amultimedia communication service, intelligent call routing based on the character-istics of a Web page that a customer is watching, synchronized pop-up of a Webpage that a customer is watching, mixed initiative control, Internet phone service,and so on. However, such Internet- based services are not available yet. Most cur-rent call centers are still second generation, integrating computer technology to-ward the third generation.

2.2 CTI-Based Call Center

CTI is literally the seamless integration or merging of computers and telephone sys-tems [1]. CTI is also defined as a technology platform that merges voice and dataservices at the functional level to add tangible benefits to business applications. Toachieve customers’ high satisfaction, many call centers today adopt CTI technologyinto PSTN as depicted in Figure 1.

Whenacustomerphonesacallcenter, theCTI-basedcallcenterprovidestheinter-active voice response service at first, through an automated machine voice services.Theotherservicesof thecurrentCTI-basedcallcentercanbesummarizedasfollows:

• Intelligent call routing: PBX monitors the status of all human agents—whetherthe human agent’s line is busy, and daily workload of each human agent. When acustomer wants to be connected with a human agent, PBX selects the most appro-priate agent based on agent status.

• Call/data synchronization (screen pop-up): When a customer’s call is connectedwith a human agent, the customer’s information pops up on the agent’s personalcomputer (PC) monitor.

MULTIMEDIA CALL CENTER FOR INTERNET COMMERCE 229

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

• Call/screen transfer: In case a human agent transfers the call to another agent,the customer’s information can be transferred.

• Expected waiting time: In case all human agents are busy, PBX provides a mes-sage of the expected waiting time to the customer, and asks whether he or shewants to wait. If the customer wants to receive a return call, PBX calls the customerwhen an appropriate human agent is available.

• Predictive dialing: In the outbound service, the CTI server with PBX calls cus-tomers automatically with a target call list. If a customer’s line is busy or if a cus-tomer does not answer, the next number on the list is dialed automatically. When acall is successfully connected, the connection signal is sent to a relevant humanagent displaying information about the connected customer and a talk script is de-livered.

2.3 Internet Extensions of CTI-Based Call Center

Because the quality of Internet telephony is not satisfactory yet, the Web-aware callcenter [2] that combines the Web service with the PSTN-based telephone service hasbeen adopted. When a customer’s request comes through the Web, the Web-awarecall center provides intelligent call routing, a pop-up screen of customer’s informa-tion, Web page pop-up that the customer is viewing simultaneously, and a call backservice. In the Web-aware call center, the customer is connected through the callback service on PSTN. However, the Web-aware call center does not have theInternet multimedia dialogue service and the initiative control function. These willbe described in the Internet-based call center scenario in Section 3.

Lucent Technologies announced a model of a call center for the Internet environ-ment in 1996 and has been working with a teleservicing company to design and in-

230 MOON, LEE, LEE

C ust o m erP S T N

P riva te B ran ch Ex chang e (P B X )

In tera ctiv e V oice R esp ons e

Serv er

C T I Server

F A X Server

H um a n

A g ents

D B S erv er

C us to m er

In fo rm a tio n

P rod ucts

In fo rm a tio n

Figure 1. Architecture of computer telephony integration-based call center. PSTN = publicswitched telephone network; CTI = computer telephony integration; DB = database.

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

stall its first operational call center for the Internet in Europe [3]. This system offersthree significant capabilities. First, uninterrupted Web voice and data can be deliv-ered over a single, regular phone line. The other current call centers for Internet re-quire a customer to have a second line or a high-speed circuit for uninterruptedWeb voice and data service. The Web customer without this extra facility must dis-connect from the Web session to free the line and wait for a call back from a humanagent.

Second, it allows a human agent to lead a customer on a tour of the Web pagesbased on input from the customers. This is called escorted browsing. The agent se-lects the pages from the business Web site and loads the pages on the customer’sPC. This capability improves the human agent’s efficiency and aids the agent inclosing the sale and/or completing the call.

Third, a call back service is provided for a customer. After a request through theWeb page, the customer waits for the return call from a human agent. The requestis routed to an agent with the proper skill, and the customer communicates withthe agent via voice or text chat over the Internet.

Although these two Internet extensions of the CTI-based call centers are impor-tant, they are still CTI-based call centers. Because their infrastructure is the same asa CTI-based call center and they still need to employ the expensive PBX, they areclassified as the third-generation call centers that are CTI and PSTN based.

2.4 Call Center As Human-Embedded Customer Servicein Internet Commerce

In this article, we view the Internet-based call center—the fourth generation—asan important tool for customer service in Internet commerce. The customer ser-vice for Internet commerce can be classified into two dimensions: type of interac-tion and level of automation (see Table 1). According to this taxonomy, anInternet-based call center can be positioned as real-time human-embedded cus-tomer service.

Because electronic commerce is becoming popular, many research and develop-ment projects have been devoted to automating electronic commerce activities. In-novations include automated catalogs, automated negotiation [4, 5], automatedpurchasing support [6, 7], automated customer query answering service [8], andcomparison shopping [9]. Automated systems can be further classified asdatabased, case-based [9, 10], knowledge-based [6], and/or agent-based system

MULTIMEDIA CALL CENTER FOR INTERNET COMMERCE 231

Table 1Classification of Customer Service by Type of Interaction and Level of Automation

Level of Automation

Type of Interaction Human Embedded Fully Automated

Real-time Internet-based call center;text-based chatting

Salesman expert system; search engine;comparison shopping service; productrecommendation service

Non-real-time Query answering emailservice

Some subscription, authentication, andsales processing

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

[5]. As shown in Figure 2, the level of automation has a technically possible limit(Point A in Figure 2) at a particular time and the cost of automating high-level ser-vices sometimes can be higher than the benefit from the automation (Range B inFigure 2). For higher level customer service where the marginal cost of the hu-man-embedded service is lower than that of automated service (Point C in Figure2), we need human-embedded services such as e-mail response and call centers.These human-embedded services will continue to exist until we can have cost-ef-fective artificial intelligence tools such as real-time speaker-independent speechrecognition capability, real-time natural language understanding capability, and ashared electronic commerce ontology between customers and businesses.

3. CALL CENTERS IN INTERNET SHOPPING MALLS

Current Internet shopping malls have many automated services including auto-mated catalogs, order processing, comparison shopping, and so on. However, cus-tomers may sometimes need a human agent for further information.

3.1 An Internet-Based Call Center Service Scenario

Suppose a company, Comm&Com, which has implemented a commercial Internetshopping mall, Cyber-Com. Assume that Cyber-Com deals with computers and pe-ripheral equipment. A customer, Tom, enters Cyber-Com and navigates to its homepages to purchase a Pentium notebook computer. Tom finds a model but cannot un-derstand the complicated specifications. Therefore, he wants to know more aboutthe model, and also compare it to other models as well. He can click the Help Me

232 MOON, LEE, LEE

Figure 2. Cost-benefit curve of customer service (automated vs. human embedded).

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

button on the Web page. Then, the call center recognizes Tom’s identification andthe page that he is on, and provides the Internet multimedia dialog service withoutthe use of a human agent. However, suppose a customer cannot find the anticipatedanswer from the Internet multimedia dialog service and wants to talk to a humanagent in the call center? He can click the Agent Connection button and will beconnected with a human agent. Before the connection, the agent can acquire cus-tomer information and the page that Tom is currently on. Therefore, the agent cangrasp the customer’s needs without a long dialogue. Voice dialogue can also takeplace through the Internet, and the agent may take the initiative of synchronous dis-play to lead the customer to the explanatory pages.

3.2 Features of Internet-Based Multimedia Call Center

AsInternet-basedelectroniccommerceevolvesandcustomerrequestsbecomemorecomplicated than in the just mentioned scenario, the Internet shopping malls shouldbe equipped with more advanced call centers. In the near future, the network band-width of the Internet will be upgraded enough for full-duplex voice communication.By the year 2003, the super Internet is expected to be implemented with a bandwidthof 1 giga-bps or 1 tera-bps. This is almost advanced enough for online two-way voicecommunication through TCP/IP (Transmission Control Protocol/InternetworkingProtocol). When this occurs, the market for the PSTN-based telephone service maysignificantly shift to the Internet platform.

The call center would obviously not be PSTN-based but Internet-based andwould be able to provide interactive multimedia services—voice, text, and imagewith synchronized browsing capability. In this article, we designed anInternet-based multimedia call center, IMC. The essential functions of theInternet-based multimedia call center are summarized in Table 2.

In Table 3, IMC is compared with existing call centers such as the CTI-based callcenter and the Internet extension of the CTI-based call center. The features of IMC,compared with CTI-based call centers, can be summarized as follows:

• IMC provides not only intelligent call routing and a pop-up screen of custom-ers’ information—both currently possible in CTI-based call centers—but also anInternet phone, a pop-up Web page, and initiative control.

• The Internet phone becomes the main tool for voice communication. The cus-tomer can talk to a human agent while navigating the Web with only one Internet line.

• An automated Internet multimedia dialogue service is provided without ahuman agent’s help. The productivity of a call center can be maximized by mini-mizing the number of calls involving human agents without sacrificing customersatisfaction [11]. Multimedia service, including voice, image, and text, can alleviatethe human agent’s workload and increase customer satisfaction.

• The most suitable human agent for the customer’s needs can be selected basedon the type of Web page that the customer is viewing as well as the customer’sidentification.

• The cost of installation or upgrade can be reduced. The Internet extensions ofCTI-based call centers require existing equipment of CTI-based call centers andother machines for the Internet. However, IMC is independent of CTI technology

MULTIMEDIA CALL CENTER FOR INTERNET COMMERCE 233

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

and a completely redesigned model based on a pure Internet platform. Therefore,the expensive PBX is not necessary.

4. DESIGN OF AN IMC

4.1 Architecture of IMC

IMC is logically divided into two systems: the Internet multimedia dialogue systemand the human agent assisting system. The Internet multimedia dialogue systemsupports multimedia interaction between customers and IMC without human in-tervention, and the human agent assisting system supports human agents in com-municating with customers.

234 MOON, LEE, LEE

Table 2Function of Internet-Based Multimedia Call Center

Function Description

Internet multimedia dialogueservice

Internet version of interactive voice response in the CTI-basedcall center; multimedia service is possible

Intelligent call routing Calls are routed to a human agent based on both the agents’status and the Web page that the customer is watching

Customer information screenpop-up

Same as that of CTI-based call center

Web page pop-up and theinitiative control

Human agent can view the same page that the customer isviewing and take the initiative of synchronous display tolead the customer to a certain page, which is the solution ofcustomer’s request

Voice communication by Internetphone

Both human agent and customer talk each other throughInternet phone

Call passing Same as call/screen transferring function in the CTI-based callcenter

Note. CTI = computer telephony integration.

Table 3Comparison of Call Centers

FeaturesCTI-Based Call

CenterInternet Extension of

CTI-Based Call CenterInternet-Based

Multimedia Call Center

Customer Phone user Internet user Internet userPlatform PSTN PSTN/Internet InternetVoice

communicationVia PSTN

telephoneVia PSTN/Internet

phone (call back)Via Internet phone

PBX Necessary Necessary Not necessaryCall routing Phone number

basedWeb page and

customer basedWeb page and

customer basedWeb page

synchronizationImpossible Possible Possible

Initiative control Impossible Impossible PossibleInternet multimedia

dialogue systemVoice only Not prepared Multimedia

Note. CTI = computer telephony integration; PSTN = public switched telephone network; PBX =private branch exchange.

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

IMC physically consists of a Web server, a human agent assisting server, anInternet multimedia dialogue server, an Internet phone server, and human agentsas depicted in Figure 3. The human agent assisting server is in charge of intelligentcall routing and the pop-up screen with the customer’s information. The Internetmultimedia dialogue server is an Internet version of the interactive voice responseserver in CTI-based call centers. The Internet phone server supports the voice com-munication between the customers and the human agents over the Internet. Hu-man agents are equipped with PCs and Internet phone equipment. The processdiagram of IMC is depicted in Figure 4 and the details are described in the follow-ing sections.

4.2 The Internet Multimedia Dialogue System

The goal of the Internet multimedia dialogue system is to provide the relevant infor-mation that a customer wants on the Internet shopping mall without involving thehuman agents. The service of Internet multimedia dialogue system can be designedas follows:

• User identification (ID) and uniform resource locator (URL) recognition: When aregistered customer clicks the Help Me button, the Web server recognizes the cus-tomer’s user ID and the URL of the Web page that the customer is viewing. Theuser ID and the URL are used for intelligent individualized services.

• Interactive multimedia service: Not only can the automated voice services beprovided, but the Web pages relevant to the voice services can be provided as well.Therefore, the customers can hear the voice response via the Internet phone whilelooking at the display of the relevant Web pages.

MULTIMEDIA CALL CENTER FOR INTERNET COMMERCE 235

Figure 3. Physical architecture of an Internet-based multimedia call center. IMC =Internet-based multimedia call center; URL = uniform resource locator.

Web server

Human Agents

Internet

Human agent

assisting server

IMC

Customer

URL-Agent Group Database

Agent Status Database

Purchase History Database

Customer Profile Database

Multimedia Dialogue Content

Database

Product Database

Consultation History Database

Internet phone server

Internet multimedia

dialogue server

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

• Menu driven inquiry: Customers can ask questions by clicking the menu on theWeb page.

• Three-dimensional (3D) presentation: A 3D display of products or services ispossible by motion pictures and virtual reality technology to enhance customercomprehension.

4.3 The Human Agent Assisting System

When the customer cannot get a satisfactory solution from the Internet multimediadialogue system, he/she may want the human agent service, which can be initiatedby the customer’s clicking the Agent Connection button. The human agent as-sisting system identifies the customer’s user ID, the URL, and the dialogue data ex-changed between the customer and Internet multimedia dialogue system. The hu-man agent assisting system is composed of eight steps. Let us assume the humanagent in our scenario is Judy.

Step 1. Availability notification of the agent: The agent, Judy, clicks to signal heravailability to the human agent assisting server. The customer information displayon her PC is on standby until the customer, Tom, calls.

Step 2. Connection request to agent: If Tom cannot acquire what he wants from theInternet multimedia dialogue system, he may want to talk to a human agent rightaway. By clicking the Agent Connection button on the Web page, Tom’s ID andthe URL of the Web page that Tom is on is forwarded to the human agent assistingsystem.

236 MOON, LEE, LEE

CustomerCustomer Human AgentHuman Agent

ServersServers

Browser IMC-ConferenceCustomer Information

Display IMC-ConferenceBrowser

Internet Shopping

Mall Log-in

“Help Me” Click

Menu Selection

“Agent Connection”

Click

Internet Phone

Communication

Collaborative

Browsing

Availability

Notification

Customer Information

Screen Pop

Call Passing

Internet Phone

Communication

Collaborative

Browsing

Collaborative

Browsing

Internet Multimedia Dialogue Server

User-id/URL Recognition

Web Server Internet Phone Server

Human Agent

Assisting Server

Internet Shopping

Mall

Agent Group Selection

Web Page Pop

Internet Phone

Communication

Agent SelectionInternet Multimedia Dialogue Service

Menu Driven Inquiry

3-Dimensional Presentation

Customer Information

Screen Pop

Internet Multimedia Dialogue System Human Agent Assisting System

Call Passing

Figure 4. Process diagram of an Internet-based multimedia call center (IMC). ID = user iden-tification; URL = uniform resource locator.

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

Step 3. Agent group selection: The URL–agent group database in Figure 3 con-tains the relation between URLs of Web pages in the Internet shopping mall andrelevant professional agents. From the database, the human agent assisting serverretrieves an agent group ID that matches the current URL.

Step 4. Agent selection: The agent status database in Figure 3 maintains the infor-mation on individual agent’s availability. The human agent assisting server re-trieves an available agent and the IP address of his/her PC from the agent statusdatabase. For that, the agent group ID in Step 3 should be used as a primary key.This agent’s information will be used in Step 5.

Step 5. Customer information pop-up screen: The customer’s information is typi-cally composed of the customer’s profile, purchase history of products or services,and consultation histories. The human agent assisting server retrieves and displaysthe customer’s information on the designated agent’s PC using the customer’s ID.The agent can understand the customer before the connection. After the screenpops up, the customer information display initiates IMC–conference on the agent’sPC, which controls collaborated browsing and Internet phone communicationwith a customer.

Step 6. Internet phone communication: The response of a successful connectionwith a human agent is sent to the customer and IMC–conference on the customer’sPC is initiated, which controls collaborative browsing and Internet phone commu-nication with the agent. The agent’s IMC–conference retrieves the IP address of thecustomer’s PC from the human agent assisting server and connects to the cus-tomer’s PC through Internet phone server. Now, the customer and the agent cantalk to each other through the Internet phone.

Step 7. Web page pop-up and the initiative control: The agent’s IMC–conference re-trieves the URL of the Web page that the customer is viewing from the humanagent assisting server and sends the URL of the request to the Web server for re-trieval of the same Web page. The same page that the customer is viewing is dis-played on the browser of the agent’s PC and collaborative browsing is started.Either the customer or the agent may have the first initiative. Therefore, when onemoves Web pages, the other can see the Web pages while talking over the Internet.If necessary, the initiative can be shifted to the other side.

Step 8. Call transferring: When transferring a call to another agent is necessary,the agent makes a request to the human agent assisting server, and the dialogue be-tween the customer and the old agent is transferred to a new agent. The humanagent assisting server brings up the customer’s information on the new agent’sscreen (Step 5), and both the customer and the new agent communicate through theInternet phone (Step 6). Through the Web page pop-up and the initiative controlfunction (Step 7), the new agent can browse the same Web page with the customer.

5. IMPLEMENTATION OF A PROTOTYPE SYSTEM: IMC–1

Based on the design of IMC, we have implemented a prototype system calledIMC–1. For the purpose of easy and fast implementation of Internet phone commu-nication and collaborative browsing, we adopted Netscape Conference in NetscapeCommunicator. We have developed the Internet multimedia dialogue service of

MULTIMEDIA CALL CENTER FOR INTERNET COMMERCE 237

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

IMC and Steps 1 to 5 in Section 4.3. Steps 6 and 7 were implemented using NetscapeConference.

The overall process of IMC–1 is similar to IMC. IMC–1 has the same functions asIMC except the call transferring function (Step 8 in Section 4.3). The initial screenfor the human agent, Judy, is Figure 5. On the left is her Web browser, and on theright is the customer information display.

Suppose the customer, Tom, enters his user ID and clicks the OKbutton to getinto an Internet shopping mall named Cyber-Com. Before getting into Cyber-Com,user authentication is performed. As we explained in the scenario of Section 3.1,Tom wants to buy a notebook PC and at first wants to compare products. He clicksthe Help Me button on the Web page of Cyber-Com, and he can see the menu pageof the Internet multimedia dialogue service. He can also hear the voice through theInternet phone explaining the detailed description of this page. Therefore, he caneasily understand what each menu means. In this page, Tom can be connected to ahuman agent directly by clicking the Agent Connection button.

Let us suppose Tom clicks to compare personal computers, the result is then dis-played (Figure 6). Tom is able to observe the page and listen to the recorded voice ex-planations over the Internet. However, if he cannot find the solution from theInternetmultimediadialogueservice,hemayclicktheAgent Connection button.

After clicking the Agent Connection button, Tom will see the message on thescreen that says “Your request is accepted. Please wait for a moment and follow ourdirections.” Now, Tom’s request for a human agent is accepted.

The human agent assisting system recognizes Tom’s request and selects Judy asthe most appropriate human agent through the agent group selection and agent se-lection functions. Judy can see Tom’s information automatically through customerinformation display as depicted on the right in Figure 5. Judy and Tom can startNetscape Conference through Netscape Communicator. Through the Internet

238 MOON, LEE, LEE

Figure 5. The human agent assisting system.

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

phone communication and collaborative browsing functions of Netscape Confer-ence, Judy and Tom can communicate with each other through the Internet phone,and Judy can see the same Web page that Tom is browsing as depicted on the left inFigure 5. At this time, Judy has the initiative and she can control Tom’s browser tolead him to other Web pages. The initiative may be transferred to Tom by request.By using IMC–1, Tom can get the solution easily from the call center and the callcenter can acquire higher customer satisfaction.

6. CONCLUSIONS

As more and more electronic commerce transactions occur in cyberspace, custom-ers will demand that the current PSTN-based call center services be provided on theInternet. Although current call centers provide only voice services, customers willdemand multimedia services that are possible through Internet-based call centers.Therefore, the services of the existing call centers based on CTI need to be upgradedto adapt to the new environment of electronic commerce. In addition, some modelsof Internet extensions of the CTI-based call centers are not advanced enough to meetthe needs of customers on the Internet. In this article, we designed a call centernamed IMC and implemented a prototype system named IMC–1. IMC can be wellintegrated with an Internet shopping mall. It can satisfy the increasing needs of theInternet commerce customers by providing individualized high-quality multime-dia services. IMC can be positioned as a real-time human-embedded system thatcan provide high-quality services cost-effectively for Internet commerce.

We see future research issues related with call centers for Internet commerce.Data mining in call centers is one example [11]. In addition, another important is-sue is business process re-engineering in adopting a call center in an organization.Intelligent outbound call management in Internet-based call centers also deservesfurther research.

MULTIMEDIA CALL CENTER FOR INTERNET COMMERCE 239

Figure 6. Internet multimedia dialogue service.

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014

REFERENCES

[1] PassageWay Direct Connection Team, Computer Telephony Integration. Murray Hill, NJ: LucentTechnologies, 1996.

[2] E. Margulies, Understanding the Voice-Enabled Internet. New York: Flatiron, 1996.[3] Lucent Technologies, “Internet call center,” Author, Murray Hill, NJ, White Paper, 1996.[4] J. R. Oliver, “A machine learning approach to automated negotiation and prospects for electronic

commerce,” Journal of Management Information Systems, vol. 13, no. 3, pp. 83–112, 1996.[5] A. Chavez and P. Maes, “Kasbah: An agent marketplace for buying and selling goods,” in Proc. of

the 1st Int. Conf. on the Practical Application of Intelligent Agents and Multi-Agent Technology, 1996, pp.75–90.

[6] S. K. Lee, J. K. Lee, and K. J. Lee, “Customized purchase supporting expert system: UNIK–SES,”Expert Systems With Applications, vol. 11, no. 4, pp. 431–441, 1996.

[7] G. Linden, S. Hanks, and N. Lesh, “Interactive assessment of user preference models: The auto-mated travel assistant,” in Proc. of the 6th Int. Conf. on User Modeling, 1997, pp. 67–78.

[8] R. Burke, K. Hammond, V. Kulyukin, S. Lytinen, N. Tomuro, and S. Schoenberg, “Question an-swering from frequently asked question files,” AI Magazine, vol. 18, no. 2, pp. 57–66, 1997.

[9] J. K. Lee, Y. U. Song, J. W. Lee, “A comparison shopping architecture over multiple malls: Themeta-malls architecture,” in Proc. of the ICEC 98, 1998, pp. 149–154.

[10] I. Vollrath, W. Wilke, and R. Bergmann, “Case-based reasoning support for online catalog sales,”IEEE Internet Computing, vol. 2, no. 4. pp. 47–54, 1998.

[11] P. Xia, “Knowledge discovery in integrated call centers: A framework for effective cus-tomer-driven marketing,” in Proc. of the 3rd Int. Conf. on Knowledge Discovery in Databases and DataMining, 1997, pp. 279–282.

240 MOON, LEE, LEE

Dow

nloa

ded

by [

UT

SA L

ibra

ries

] at

07:

37 0

7 O

ctob

er 2

014