Applications of Expert Systems in Human Resource Management

Beginning of article

It is common knowledge that computers can offer the professional worker myriad benefits. Similarly well known is the fact that the misapplication of computer technology can cause much grief. Correctly ascertaining the suitability of computers in a particular situation is thus an important part of planning. The question of suitability is particularly difficult to answer when the technology is new or unfamiliar to those making the decision. To many human resource managers, the expert system (ES) -- a computer program designed to answer questions as would a human expert -- is an example of unfamiliar technology. The purpose of this article is to inductively develop a conceptual framework from a review of ES applications and demonstrate its usefulness with respect to ES technology and the tasks commonly faced in the management of human resources.

This article will begin by providing an example of interaction with a commercially-available application. Next will be the identification and definition of variables that affect the use of ES technology in all fields. Human resource management (HRM) will then be specifically addressed in terms of its four main decision areas: the acquisition, development, rewarding, and maintenance of human resources (Greenlaw & Kohl, 1986). Specific software applications, some in operation, some only conjectural, will be used as examples of the concepts presented in this article. Finally, the conclusion will present general observations regarding the applicability of ES technology to HRM.

Example of User Interaction with an Expert System

Because many readers may not have any practical experience with ES technology, this article presents an example of an ES application from the user's point of view. The application discussed makes use of a graphical user interface. Such interfaces often make it easier to learn an application, as movements of a mouse or the keyboard arrow keys can be substituted for textual commands.

This application is Policies Now!, by KnowledgePoint (1994). It is designed to assist in the creation of an employee handbook. The software is organized around the areas of concern that are typically addressed in a personnel policy statement or handbook. For example, the user could use the keyboard arrow keys to navigate to the "Leaves of Absence" section. By placing the cursor on "Personal Leave" and pressing the ENTER key, the user could choose to answer questions about the company's standard practices regarding a personal leave of absence. The software would offer some basic information regarding the importance of a personal leave policy, and would then begin a dialogue with the user to determine the company's policies. A partial example of this dialogue is shown below:

Figure 1
(Adapted from Policies Now!)

STATUS    POLICY SELECTION                         EFFECTIVE/REVISION

          501 Safety [P]
          502 Work Schedules
          504 Use of Phone & Mail Systems

ACC       505 Smoking [P]                          12-01-93 12-01-94
          506 Rest & Meal Periods

ACC       507 Overtime [P]                         12-01-93 12-01-94
          508 Use of Equipment & Vehicles

ACC       510 Emergency Closings                   12-01-93 12-0 1-94

PEN       512 Business Travel Expenses             12-01-93 12-01-94
PEN       514 Visitors in the Workplace            12-01-93 12-01-94

          LEAVES OF ABSENCE
          601 Medical Leave [P]
ACC       602 Family Leave                         12-01-93 12-01-94
          603 Personal Leave [P]
          604 Educational Leave
          605 Military Leave [P]

PEN       607 Pregnancy--Related Absenses [P]      12-01-93 12-01-94

System: "Eligibility for personal leave is commonly restricted to certain employee classifications (for example, regular full-time employees). You are free to define eligibility at your discretion as long as the guidelines are not based on discriminatory criteria."

The user is then directed to select among the various employee classifications. These classifications would have been developed earlier to fit the company's culture.

User: {Regular full-time employees} and {Regular part-time employees}

System: "What is the maximum number of days you allow for a personal leave of absence?"

User: {30} calendar days every {3} years.

System: "Do you give consideration to requests for extensions to the maximum period for a personal leave of absence?"

User: {Yes}

System: "What is the maximum number of calendar days you will allow for an extension to the initial period of personal leaves of absence?

"Enter the number of calendar days:"

User: {20}

System: "Do you require employees in an eligible employment category to complete a waiting period before requesting personal leave?"

User: {Yes}

System: "How many days of service in an eligible classification do you require before employees can request personal leaves of absence?"

"Enter the number of calendar days:"

User: {90}

The dialogue would continue until the system had gleaned all necessary information from the user. After this, the ES would be able to write the appropriate text to convey the gathered information in a readable format. The user would be able to edit this output if further polishing is desired.

Variables Affecting the Applicability of Expert System Technology

The variables that affect the degree to which an ES can be used in a given field may be categorized into three main groups: 1) Environmental variables which concern the setting in which the ES will be expected to operate; 2) User variables depending on the person using the ES at a particular moment; 3) Feasibility variables which can cause a theoretically sound ES to be unacceptable for a particular company.

Environmental

There are at least three important environmental variables. Stability is one of these. Environments which experience constant, rapid change are generally not very good candidates for ES solutions, as programmers may be unable to modify the ES quickly enough to keep up with the changes.

Structuredness is a second environmental variable. In a highly structured system, the relationship between various components is well-defined, even when the details of each component are subject to change. For example, a company may use many different techniques to recruit for its sales force, but the general strategy of attracting qualified people remains constant. Successful ES application generally requires some degree of structure. The third environmental variable is ambiguity. The use of eye contact and voice inflection by a job applicant, the extra effort …