Anecdotal reports and recent reviews assert that expert systems are potentially useful decision aids in human resource management. This study examines the effects of an expert system designed to aid employees when they make their choices in a flexible bellcfit program. A four group quasi-field experimental design is used to examine the relative effects of the expert system compared to a conventional spreadsheet decision aid. Eighty employees at an NCR-AT&T facility were randomly selected and assigned to the groups. Employees using the expert system expressed greater benefits satisfaction compared to those using the spreadsheet aid. The spreadsheet did not have any effect on employees' decisions. When the benefit choices recommended by the expert system differed from the employees' current choices, employees are more likely to change their choices. Consequently, the expert system is likely to affect employees' decisions. Implications are discussed and future research needs are suggested.