![]() |
|||
![]()
|
![]() |
![]() |
![]() |
3.3. BUSINESS APPLICATION AGENTSIntelligent agent technology is being used in a wide variety of applications related to business operations. Agent-based systems are in use in customer service, human resources, manufacturing, sales, and several other areas. Business application of data-mining agents is discussed in Section 3.5. Intelligent agents are being used as automated customer service agents, providing information to customers without the need for human intervention. For example, intelligent agents are used in systems to provide banking information to bank customers and information on the status of shipments to customers of shipping companies. Customer service systems such as these are less expensive than human customer service personnel. They are often faster in providing standard requested information, and they allow the human customer service personnel to be used to provide answers to more difficult or unusual requests. Another major advantage of the agent systems is that they can offer some degree of intelligent service to customers 24 hours a day, 7 days a week.
Figure 1 illustrates a customer support agent that carries out a dialog with users on the Web to solve routine hardware and software problems. For example, a user may be having trouble with a printer or a word-processing application. Instead of calling a help desk about the problem, the user could have a dialog with the support agent on the Web. The support agent becomes the first line of response to customer problems. The agent is meant to only address relatively simple problems that are easy for the user to specify and can be resolved without complex steps. The agent starts the dialog by asking the user a question. The set of possible answers to the question are also provided. Since the goal is to keep the dialog simple, questions have three to four possible answers. For example, the question may ask the type of printer being used. The user selects one response from the available choices. The user's response tells the agent a single fact about the problem. The user can also indicate that the question is irrelevant to the problem. The response to each question is added to the representation of the user's problem being assembled by the agent. The agent's knowledge base contains additional knowledge about the class of problems it can resolve. This domain knowledge allows the agent to infer additional knowledge about the user's problem. For example, knowing what printer the user has, the agent could infer that this type of printer has been having ribbon problems. Then, the agent could ask the user to check if the ribbon is properly installed on the printer. The agent's representation of problems consists of a set of problem attributes with specified values. The solution to a problem is represented as a text file describing the solution. Essentially, the purpose of the dialog is to enable the agent to understand the user's problem. This understanding is represented as the values of attributes that underlie the questions the user is asked. When the user's problem matches the stored representation of a problem in the knowledge base, the agent retrieves the corresponding solution file and provides the contents of the file to the user. Agents are also being used to provide systems for business operations, such as human resources departments. For example, agents are an important enabler of a paperless wage review system for a division of Hewlett-Packard. Each of over 10,000 employees is reviewed quarterly. Intelligent agents in this review system insure that the wage review process for all of these employees is performed in a timely manner. The agents automatically initiate the wage review process on the appropriate date for each employee, sending the wage review forms to the manager. Then, they monitor the progress of the process and insure timeliness. Employee benefits systems have been developed to automate other human resources tasks. For example, such systems provide benefits information to employees, or make available an employment verification service to banks and other external companies that need to verify employee status and title. Agents are in use in manufacturing applications. They are being used for several different purposes, for example, to plan, verify, or optimize production schedules, as well as to gather and integrate manufacturing data. In sales, agent-enabled sales systems have been developed to link a company's sales force with its manufacturing and distribution systems. When such a system receives a request for delivery of a certain quantity of a particular product, it can check the availability of the product or its raw materials, the availability of delivery capacity, etc., and give a real-time response. Some systems utilize constraint-based reasoning to allocate or suggest allocation of company manufacturing and distribution resources to optimize delivery schedules and costs. A key term being used in the business application domain is online analytic processing (OLAP). Comshare has a program called Robot for OLAP that searches OLAP databases for trends and patterns using multidimensional views of the data. This system is used by the Hertz car rental company to analyze pricing moves by other car rental companies. Allied Signal is also using the Comshare OLAP software to provide sector level business information. Previously, they only had business unit level information and did not have rework costs available for cost of quality computations. Comshare also has robots to monitor news feeds, stock quotes, and for internal Lotus Notes databases. This capability illustrates the evolution of executive business information needs to encompass external information and data stored in in-house intranets.
|
![]() |
|
Use of this site is subject certain Terms & Conditions. Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details. |