6
AbstractInternet has opened new doors between a company and its customers as companies are preparing to shift their focus from transaction to relationships. Today it is widely acknowledged that how a company treats its customers goes a long way in determining its future profitability, and companies are making bigger and bigger investments on it. Firms that develop strategies and technologies for implementing customer relationship management, with the most profitable customers, become winners in the intensely competitive, dynamic markets of the digital economy. Intelligent agents are a rapidly developing area of research. It is an emerging technology, which has its roots firmly in AI research. With the proliferation and evolution of the Internet, it has also increasingly emerged as a means to achieve customer satisfaction and delight by the way of extensive e-CRM. However, it is not yet very clear how agents can be used in the domain of e-CRM. In the first part, this paper provides an overview of intelligent agents. The second part presents the ways in which enhanced customer attainment and retention can be achieved with e-CRM. Finally an integrated model is proposed wherein we suggest how intelligent agents can be used as powerful tool to achieve e-CRM over the Internet. Index TermsIntelligent agents, CRM, e-CRM, bots, e- Business. I. INTRODUCTION It is called customer management, customer information systems, customer value management, customer care and sometimes customer centricity or customer-centric management. But clearly, now, the term Customer Relationship Management (CRM) has overtaken the market. CRM is a business strategy to select and manage the most valuable customers. CRM is a customer-centric business philosophy and culture that support effective marketing, sales, and service processes. The ideas behind CRM are not new. CRM in a sense is the way the corner grocer used to treat his customers. What is new is that we now can do it on an "industrialized" basis for tens of thousands, even millions of customers. That's what is entirely new. We have gone back to the old way of doing business, a customer at a time but for millions of customers. That is CRM today. The sheer endlessness information available through the Internet, which at first glance looks like its major strength, is Manuscript received April 11, 2008 Authors are with XLRI, Jamshedpur, INDIA (phone: 91-657-3983150; fax: 91-657-2227814; e-mail: [email protected] and [email protected] . at the same time one of its major weaknesses. Many attempts have been made to provide Internet and Internet users with tools that aid them in finding valuable information in the Web. Most of these attempts have resulted in content-based search engines like Alta Vista, Excite, Infoseek and Google. However, these tools are geared towards finding as much information that is related to the keyword for which a search is performed. This means that a search on, e.g., the keyword “e-commerce” on the WWW leads to 161,000,000 answers. A way to filter these answers is by using user profiles. A user profile is usually more than a set of keywords in which the user is interested. These keywords are “added” to any search of the user to filter out only those items that are of interest for the user. The user profiles can be static or dynamic. In the fist case, they are defined once and can only be changed by an explicit update from the user. The dynamic user profiles use some type of learning algorithm to update the profile, either through explicit feedback of the user or through observing the behaviour of the user [7]. A solution to the problem of getting the right information for the users can be sought by the use of Intelligent Agents. An agent is an autonomous and adaptive entity that is situated in some environment and performs purposefully in collaboration with other agents or human beings [4]. The article is not about the development of a new agent theory. Existing theories and agent libraries have been used as far as possible. The main contribution lies in the combination of agents that is used and the roles they fulfil in the e-CRM. In subsequent sections, the different types of intelligent agents are described which play a role in the model. After the description of this model, few practical considerations are described which have to be taken into account when the model is to be implemented. II. INTELLIGENT SOFTWARE AGENT Agent software is an emerging technology, which promises to be many things to many people, however the technology is still in an embryonic stage. Despite this, the range of organizations and disciplines researching and pursuing agent technology is broad. The more recent notable applications for agents include e- commerce, Web marketing and Internet search agents. What these applications have in common of course is the Web. It seems that with the many types of Web-based software being developed, the Web is providing a good architectural framework for the development of agents. This is because the Web facilitates many of the characteristics of agent software Implementing e-CRM using Intelligent Agents on the Internet Ashis K. Pani and Pingali Venugopal 978-1-4244-1672-1/08/$25.00 ©2008 IEEE.

[IEEE 2008 International Conference on Service Systems and Service Management (ICSSSM 2008) - Melbourne, Australia (2008.06.30-2008.07.2)] 2008 International Conference on Service

  • Upload
    pingali

  • View
    217

  • Download
    1

Embed Size (px)

Citation preview

Abstract—Internet has opened new doors between a company and its customers as companies are preparing to shift their focus from transaction to relationships. Today it is widely acknowledged that how a company treats its customers goes a long way in determining its future profitability, and companies are making bigger and bigger investments on it. Firms that develop strategies and technologies for implementing customer relationship management, with the most profitable customers, become winners in the intensely competitive, dynamic markets of the digital economy.

Intelligent agents are a rapidly developing area of research. It is an emerging technology, which has its roots firmly in AI research. With the proliferation and evolution of the Internet, it has also increasingly emerged as a means to achieve customer satisfaction and delight by the way of extensive e-CRM. However, it is not yet very clear how agents can be used in the domain of e-CRM. In the first part, this paper provides an overview of intelligent agents. The second part presents the ways in which enhanced customer attainment and retention can be achieved with e-CRM. Finally an integrated model is proposed wherein we suggest how intelligent agents can be used as powerful tool to achieve e-CRM over the Internet.

Index Terms— Intelligent agents, CRM, e-CRM, bots, e-Business.

I. INTRODUCTION It is called customer management, customer information

systems, customer value management, customer care and sometimes customer centricity or customer-centric management. But clearly, now, the term Customer Relationship Management (CRM) has overtaken the market. CRM is a business strategy to select and manage the most valuable customers. CRM is a customer-centric business philosophy and culture that support effective marketing, sales, and service processes.

The ideas behind CRM are not new. CRM in a sense is the way the corner grocer used to treat his customers. What is new is that we now can do it on an "industrialized" basis for tens of thousands, even millions of customers. That's what is entirely new. We have gone back to the old way of doing business, a customer at a time but for millions of customers. That is CRM today.

The sheer endlessness information available through the Internet, which at first glance looks like its major strength, is

Manuscript received April 11, 2008 Authors are with XLRI, Jamshedpur, INDIA (phone: 91-657-3983150; fax:

91-657-2227814; e-mail: [email protected] and [email protected].

at the same time one of its major weaknesses. Many attempts have been made to provide Internet and

Internet users with tools that aid them in finding valuable information in the Web. Most of these attempts have resulted in content-based search engines like Alta Vista, Excite, Infoseek and Google. However, these tools are geared towards finding as much information that is related to the keyword for which a search is performed. This means that a search on, e.g., the keyword “e-commerce” on the WWW leads to 161,000,000 answers. A way to filter these answers is by using user profiles. A user profile is usually more than a set of keywords in which the user is interested. These keywords are “added” to any search of the user to filter out only those items that are of interest for the user. The user profiles can be static or dynamic. In the fist case, they are defined once and can only be changed by an explicit update from the user. The dynamic user profiles use some type of learning algorithm to update the profile, either through explicit feedback of the user or through observing the behaviour of the user [7].

A solution to the problem of getting the right information for the users can be sought by the use of Intelligent Agents. An agent is an autonomous and adaptive entity that is situated in some environment and performs purposefully in collaboration with other agents or human beings [4].

The article is not about the development of a new agent theory. Existing theories and agent libraries have been used as far as possible. The main contribution lies in the combination of agents that is used and the roles they fulfil in the e-CRM. In subsequent sections, the different types of intelligent agents are described which play a role in the model. After the description of this model, few practical considerations are described which have to be taken into account when the model is to be implemented.

II. INTELLIGENT SOFTWARE AGENT Agent software is an emerging technology, which promises

to be many things to many people, however the technology is still in an embryonic stage. Despite this, the range of organizations and disciplines researching and pursuing agent technology is broad.

The more recent notable applications for agents include e-commerce, Web marketing and Internet search agents. What these applications have in common of course is the Web. It seems that with the many types of Web-based software being developed, the Web is providing a good architectural framework for the development of agents. This is because the Web facilitates many of the characteristics of agent software

Implementing e-CRM using Intelligent Agents on the Internet

Ashis K. Pani and Pingali Venugopal

978-1-4244-1672-1/08/$25.00 ©2008 IEEE.

such as mobility and communication. Agent technology has emerged from the field of AI

research. The term 'agent' can be thought of as an umbrella under which many software applications may fall, but is in danger of becoming a noise term due to over use [10]

There are a number of classifications schemes that can be used to type-cast existing agents, for example mobile or static, deliberative or reactive, but Nwana [8] classifies agents according to primary attributes which agents should exhibit. The three primary attributes are cooperation, learning and autonomy. Nwana uses the diagram in Figure-1 to derive four types of agents, Collaborative, Interface, Collaborative Learning and Smart agents. However, Nwana recognizes that the categories in the diagram are not definitive and agents can also be classified by their roles, and so adds to that list Mobile, Infortmation/lnternet, Reactive and Hybrid agents.

Collaborative Smart Agents Learning Agents Collaborative Interface Agents Agents

Figure-1: Nwana’s classification.

Wooldridge [10] takes a more formal approach to the definition of agent, falling back to the more specific meanings from AI researchers. However, he notes that as the AI community cannot agree on the question of What is Intelligence? A less formal definition may be needed to include many software applications being developed by researchers in related fields. To this end, Wooldridge [10] introduces the notions of weak and strong agency.

In the weak notion of agency, the most general way in which the term agent is used, is to denote a hardware or (more usually) software-based computer system that enjoys the following properties:

⎯ autonomy: agents operate without the direct intervention of humans or others, and have some kind of control over their actions and internal state;

⎯ social ability: agents interact with other agents and (possibly) humans via some kind of agent communication language;

⎯ reactivity: agents perceive their environment and can respond to it in a timely fashion.

⎯ proactivity: agents do not simply act in response to their environment, they are able to exhibit goal-directed behaviour by taking the initiative;

⎯ temporal continuity: agents are continuously

running processes, not once-only computations or scripts that map a single input to a single output and then terminate;

⎯ goal orientedness: an agent is capable of handling complex, high-level tasks. The decision how such a task is best split up in smaller sub-tasks, and in which order and in which way these sub-tasks should be best performed, should be made by the agent itself.

Strong agency takes on the specific meaning form AI research, implying that agents must exhibit notions such as knowledge, belief, intention and obligation, with some researchers considering emotional characteristics as a requirement. If this definition of agent is strictly, adhered to, many software applications claiming to use agent technology would be rejected as such. Agents that fit the stronger notion of agent usually have one or more of the following characteristics:

⎯ mobility: the ability of an agent to move around an electronic network;

⎯ benevolence: is the assumption that agents do not have conflicting goals, and that every agent will therefore always try to do what is asked of it;

⎯ rationality: is the assumption that an agent will act in order to achieve its goals and will not act in such a way as to prevent its goals being achieved - at least insofar as its beliefs permit;

⎯ adaptivity: an agent should be able to adjust itself to the habits, working methods and preferences of its user;

⎯ collaboration: an agent should not unthinkingly accept (and execute) instructions, but should take into account that the human user makes mistakes, omits important information and/or provides ambiguous information.

Although no single agent possesses all these abilities, there are several prototype agents that posses quite a lot of them. At this moment no consensus has yet been reached about the relative importance (weight) of each of these characteristics in the agent as a whole. But these are the kinds of characteristics that distinguish agents from ordinary programs.

A term that is used interchangeably with intelligent agents is a bot. The word ‘bot’ is a short form for robot. A bot is a software tool for digging through data. You give a bot directions and it brings back answers. On the Web, robots have taken on a new form of life. Since all Web servers are connected, robot-like software is the perfect way to perform the methodical searches needed to find information[3].

III. ELECTRONIC CUSTOMER RELATIONSHIP MANAGEMENT (E-CRM)

At its core, e-CRM is a simple notion. It acknowledges that a fundamental change has taken place in the economy, and that for businesses to succeed and profit, they have no choice

Co-operate

Autonomous

Learn

but to learn from, and actively respond to, their customers' needs. Businesses armed with an understanding of what their customers want, and how and when they want it, can make informed decisions to drive business strategy, build brand awareness and attract and retain their most profitable customers. The relative success or failure of these efforts can now be measured and modified in real time, further elevating customer expectations. e-CRM has become a requirement not a competitive advantage [3].

The Internet has served to both accelerate the technology and bring various specialties of customer management together into a single concept, which is now being called e-CRM. e-CRM is a comprehensive approach which provides seamless integration of every area of business that touches the customer namely: marketing, sales, customer service and field support through the integration of people, process and technology, taking advantage of the revolutionary impact of the Internet. e-CRM is a management discipline concerned with how organizations can increase retention of their most profitable customers using Internet technologies and simultaneously reduce costs and increase value of interactions, thereby maximizing profits [1]. e-CRM allows marketers to leverage data from a variety of sources including sales, transactional, call center, legacy, ERP, click stream and 3rd party data and to interact with customers through web, wireless and voice. Ultimately, a business can leverage data, enabling them to satisfy customers' needs in a more complete way than ever before.

e-CRM lends a single view for customers. A complete, integrated e-CRM system is characterized by faster, automated services available online or on the desktop 24 hours a day. Typically the greatest challenge a global company faces is in accomplishing this level of services. Therefore, the focus of e-CRM is in integrating front-office and back-office activities and cross-divisional functions. The e-CRM business process can be mapped to the following integrated activities:

⎯ designing an interaction based on relevant information

⎯ personalizing every interaction ⎯ reaching the customer at the appropriate place and

time ⎯ facilitating the interaction and closing the ensuing

transaction Guidelines for a Successful e-CRM Solution The most critical feature of any e-CRM solution is the

ability to transform customer data, collected from a wide variety of sources, into the type of detailed customer information around which a company can organize its enterprise and build its customer relationships.

Customer data needs to be collected, cleaned, stored in a format that makes it easily accessible for analysis and then analysed by statisticians so that meaningful information regarding customer behaviour, trends and attitudes can be extracted.

e-CRM initiatives are often implemented with too many

business objectives in mind and managed by departments with different priorities and conflicting politics. What is most often missing is a fundamental rule of breaking down the business requirements and development plans to their component parts to focus on a single, clearly defined objective.

IV. INTELLIGENT AGENTS IN E-CRM This study believes that in future intelligent agents will

have a key role to play in CRM. As the Internet increasingly becomes a channel of conducting business and gathering information, users are likely to suffer from information overload. Consider a situation where a customer enters an on-line shopping store and does not find the required product because of the complexity of the site. By the virtue of the competitor being just-a-click away, there is a high chance of him/her defecting. Just like a offline store would have sales-representatives guiding the user, similar help would be needed in the on-line model in order to retain customers. It is here that the role of intelligent agents is envisaged. Agents can serve the purpose of designing better interactions with the customer and personalizing the service on channels like the Internet.

Figure-2: Three Cs: Company, Customer and Competitors

One of the goals of e-CRM is providing value-added

services to customers like price-comparisons and detailed product information depending on the customer requirements. Intelligent agents like shopbots and price-bots can serve these needs. This study believes that bots can play a vital role in the 3Cs as shown in Figure-2.

Since the focus is on the customer and managing customer relationships, the role of intelligent agents in the interface between the company and its competitors is not considered for the purposes of this study. The remaining two roles are investigated in the next sections. A . An Example

The agent at the user’s computer is capable of roaming around the various home appliances that are networked to the computer. Each of these home appliances is equipped with hardware to detect and diagnose a fault that may occur in the appliance. This information is than e-mailed to the respective

Competitors

Company Customers

Intelligent Agents

Competitor Information

Value Added Service

Personalized Agents

service center of the concerned appliance’s manufacturing company. Another option may be that the agent is mobile across the Internet and moves to the service center itself. As an alternative, the service centers for a company may have localized mobile agents, that roam around all the company’s appliances installed in a defined geographical location, say check the working of each appliance once a day, detect a fault, and report the maintenance requirement to the center.

B. Improving Customer Interactions A good e-CRM solution is possible if a company is able to

establish better customer interactions. Smart bots are intelligent software products that integrate computer interaction and natural language understanding to bring a human-like presence to the points of contact between a company and its customers. Some of the advantages of using bots are:

The use of natural language processing by smart bots makes it possible for customers to communicate with the company through the Internet as if they were communicating with a person—something that so far largely been missing from the Internet.

Smart bots can help customers 24 hours a day, every day of the year. Customers would appreciate receiving a prompt, friendly response to their inquiries whenever they contact companies online.

Smart bots are the closest simulation of the offline salesperson on the Internet.

Smart bots are patient and polite. They never get tired of answering the same customer questions, over and over again. They can remain calm and polite even in dealing with the most difficult customer. They can even react to the emotional content of the customer’s conversation.

Smart bots are good conversationalists. They can listen to customers and remember everything that was said. They can direct the conversation so the customers will learn what the company wants them to learn.

Smart bots are great business advisors. Customers leave behind an amazing amount of information when they engage in a conversation with a smart bot: information that can be used for highly accurate customer profiling, and to get a much better sense of the efficacy of online presence. No other customer-tracking tool comes close to offering the power of smart bots in understanding customers.

Different types of smart bots available are: Friendly natural language bot that resides on the Web site to

help visitors by processing and responding to questions using natural language.

Statistical text analyzers that extract useful information from conversations between customers and other bots.

Decision-making tools that allow seamless switching between the web site and a human customer support representative.

Intelligent e-mail auto-response systems that can generate prompt and accurate responses to e-mail inquiries, route e-mail according to content, handle Web-based forms, or even

act as an e-mail based information retrieval tool. Smart bots that send highly targeted sales and marketing

information to customers. Smart bots that act as the customer’s personal assistant and

exchange information with all other bots [3]. The smart bots employed for e-CRM can provide the

following services: Enhance customer experience by simplifying navigation

and adding a personalized, human touch to Web sites. Integrate Web sites and traditional call-center systems to

give customers fast, efficient service and support through a single channel.

Personalize customer interactions using a unified customer profile that draws information from multiple sources.

Give customers what they want when they want it: allow them to learn about the companies’ products and, to purchase from the company 24 hours a day, 7 days a week.

Reduce costs, handle peak request loads, and adjust to varying customer demand and personnel shifts by automating marketing, sales and support force.

V. THE AGENT E-CRM ARCHITECTURE The architecture of the model is shown in the figure-3. We

need four types of agents separate for all the customers and companies, while having a central customer profile agent and a product and service information agent.

The product and service information agent tries to optimize

the access to the heterogeneous company information sources. It has to maintain a kind of information model that can be matched with the customer requests and profiles. The product and service information agent also knows which company have, which product and service offerings. This can enhance the efficiency of the system considerably. In a later stage separate product and service information agents can be used for each type of company information. This will increase the scalability of the system.

The customer profile agent performs the same type of role toward the customers as the product and service information agent performs towards the companies. The customer profile agent keeps track of the profiles of the customers. The customer profile agent also acts proactively in that it will make a selection of new product and service information based on the customer profiles.

From the Figure-3 it can be seen that two types of agents, the customer agents and customer profile agents are taking care of the customer side of the system. The customer agent wants to get the right information for its customer. The customer agent can show intelligent behavior in two ways. First, it can maintain an optimal customer profile by monitoring the customer’s responses to the information that she gets. Also, the customer agent can try to optimize the presentation of this information for the customer by customer

Figure-3: The Agent e-CRM Architecture

Product and Service Information Agent

Company Company Interface Module

Query Handler

Communication Handler

Company Profile Manager

Company Agent

Customer Profile Agent

Customer Customer Interface Module

Query Handler

Communication Handler

Customer Profile Manager

Customer Agent

interface module. The system can be influenced from two sides. On the one

hand, there are requests of product and service information from the customers. On the other hand, there are updates on the company offerings. The queries from the customers lead to reactive behavior of the system in the sense that the customer agent will try to gather the best information to fulfill the query of the customer. On the other hand, the system will present the customer new product and service information when any of the relevant company’s product and service information is updated. Because this action is not triggered by the customer, it is called proactive behavior. The communication between the agents is relatively simple. The only messages that have to be communicated by an agent are request and the subsequent answer to that request.

The architecture of the customer agent consists of four

parts: the customer interface module, communication handler, query handler and customer profile manager. The customer profile manager consists of a list of product and service rating of the companies and a weighted list of terms that determine the most interesting product and service for the customer. The customer profile manager can be seen as the kind of data mining tool that abstracts interesting correlations out of the databases concerning the customer profiles. In fact, this forms the proactive part of the customer agent

Most often, the customer poses a query that is forwarded to the query handler by the customer interface module. The query handler compares the query with the current customer profile. On the basis of the query and the profile, a new query is composed, which is sent by the communication handler to the customer profile agent. After some time the answer to this query comes back from the Customer profile agent. The query handler will now filter the answer, based on the customer profile. The ratings of the companies and interests in the product and service is used to determine the final answer, which is forwarded to the customer interface module which in

turn will present the answer to the customer in a format that is based on the ratings of the customer profile.

The second type of interaction that the customer has with the customer interface module is to update his/her profile. In this case, the customer interface module opens up a dialogue screen in which the user can interact with the customer profile manager. She can add and/or delete from the current profile.

When a query arrives the company agent query handler decomposes the query. The company agent will send back the answers to the query from the their company. The query handler will collate the answer and send it back through the communication handler at one time to the customer profile agent. The only proactive behavior of the company agent consists of monitoring the company for new product and service offerings.

A last decision that should be made with respect to the

agent e-CRM architecture is where the agents should reside. For the company agents it seems logical that they reside on the server that contains or connects to the product and service information source. The product and service information agent is located at a central server. The customer Profile agent is located at the same server as the product and service information agent, but could also be located in a different server. Ideally, the customer agent resides on the PCs of the customers.

VI. LIMITATIONS AND FUTURE RESEARCH Inherent within any research are potential limitations that

affect the overall reliability of the system. With regard to the use of intelligent agents in CRM, a few limitations should be considered. The accuracy of the result of a customer profile manager that abstracts interesting correlation out of the database concerning the customer profiles. Further more, attitude towards providing information online and willingness is a concern. Future research should explore more on the use of natural language processing by smart bots.

A final limitation of the intelligent agent is the deliberate exclusion of issues such as trust, privacy and security that plays an important role in providing an individual’s willingness to provide information online.

Despite these limitations, our initial understanding of using intelligent agents in e-CRM has a good potential for real life application. More theoretical research is needed to explore the unique and distinctive use of intelligent agents in e-CRM. The bottom line is that a very little work has been explored about the use of intelligent agents in CRM so a continued research in this area is essential.

VII. CONCLUSION Though the Intelligent Agents are in a stage of exploration

and infancy, it has been seen that they can be used extensively and effectively to handle all aspects of customer interactions. The benefits and the pay-back of e-CRM projects are well worth the investments. It delivers profitable and sustainable revenue growth.

VIII. ACKNOWLEDGEMENTS

The authors would like to thank the anonymous referees for their valuable suggestions and comments.

REFERENCES

[1] S. A. Brown, Customer Relationship Management: A strategic imperative in the world of e-Business, John Wiley & Sons, Toronto, 2000.

[2] F. Dignum, “Information management at a bank using agents: Theory and practice”, Applied Artificial Intelligence, Vol. 14, pp. 677-696, 2000.

[3] O. Etzioni and D.S. Weld, “Softbot-Based Interface to the Internet”, Communication of the ACM, Vol. 37, No. 7, pp. 72-76, 1994.

[4] D. Gilbert, et al “The role of intelligent Agents in the information infrastructure”, IBM, 1995. Available on http://activist.gpl.ibm.com:81/whitepaper/ptc2.htm

[5] D. B. Lange and M. Oshima, “Dispatch your agents, shut off your machine”, Communication of the ACM, Vol. 42, No. 3, pp. 88-89, 1999.

[6] M. Ma, “Agents in e-commerce”, Communication of the ACM, Vol. 42, No. 3, pp. 78-80, 1999.

[7] P. Meas, “Agents that reduce work and information overload”, Communication of the ACM, Vol. 37, No. 7, 1994.

[8] H. Nwana, “Software Agents: An Overview”, Knowledge Engineering Review, Vol. 11, No. 3, 1996.

[9] M. Weiss, “A gentle introduction to agents and their applications”, Mitel Corp., http://www. Magma.ca/~mrw/agents/, 1999.

[10] M. Wooldridge and N. Jennings, “Intelligent Agents: Theory and Practice”, Knowledge Engineering Review, Vol. 10, No. 2, 1995.

[11] A. S. Vivacqua., “Agent for Expertise Locations”, Technical Report SS-99-03, Stanford University, 1999.