Discussant's response to "Interim report on the development of an expert system for the auditor's loan loss evaluation";

              Discussant's Response to "Interim Report on the Development of An Expert System for the Auditor's Loan Loss Evaluation"
William F. Messier, Jr.
University of Florida
It is a pleasure for me to comment on the paper by Kelly, Ribar and Willingham. As someone who has spent a major part of the last five years in expert systems research, it is good to see this technology begin to impact audit practice. I will make one caveat before I proceed. There are times in my discussion where I raise questions about or criticize this work. In those instances, please recognize that I am fully cognizant of the difficulties of doing this type of research and, more importantly, sympathetic with those difficulties.
Before discussing the specifics of the paper, I would like to make a few overall comments on this work. The research by Kelly, et al. is noteworthy for three reasons. First, it demonstrates the application of expert systems technology to an important audit problem, the assessment of loan loss reserves. Application of this technology to auditing is important because public accounting firms are facing a more competitive environment that will require audits to be conducted with the same level of effectiveness but with increased efficiency. Relatedly, the types of decisions auditors face today are more complex (e.g., EDP auditing) and require more expertise. Expert systems are intended to assist with such complex decisions.
Second, since this is a proprietary system, it is especially noteworthy that Peat Marwick is willing to share the details of the system with academics and practitioners. Until recently, many public accounting firms were unwilling to share these types of developments with the public. I make a point of this because I believe that it is important to our profession to disseminate research and that it should be a two-way street.
Finally, this paper shows that public accounting firms are willing and able to build expert systems. This realization was also brought home to me at the recent expert systems conference at the University of Southern California where Coopers & Lybrand demonstrated an expert system for deferred taxes [Shpilberg and Graham, 1986]. For that project, Coopers & Lybrand hired a full-time computer scientist to assist in developing their system. By bringing their enormous resources to bear on the problem, they were able to construct the system in approximately a year. Peat Marwick has been able to do a similar sort of thing with CFILE. In spite of the generous support from Peat Marwick's Research Opportunities in Auditing grants, most academics who are developing expert systems have faced much longer development times. This raises a question about whether academics any longer have a competitive
182