Computer Mentors Give Loan Officers a Hand

Article excerpt

Help in making commercial loan decisions has arrived in the form of expert systems software

It is not breaking news to bank managements that commercial lending today is riskier and more competitive than ever, margins are shrinking at faster paces, and lending staffs are getting younger and less experienced. In the past year or so, some U.S. banks have installed expert systems-software programs that incorporate the knowledge and judgments of experts-to combat these problems. A few have purchased comprehensive expert systems that handle each step in the commercial lending process, in an effort to reduce risk and achieve a competitive advantage.

Talk of expert systems is certainly not new in banks. It is true, however, that the growth of this technology has been slow, and bankers seem reluctant to embrace it fully. Bankers are by nature conservative. On the other hand, the technology has improved in the past two years, making it more attractive to bank managements.

Lack of a common definition. A software program might be considered an expert system by some bankers but not by others. There seems to be no common definition. Some bankers might call an analysis software program that performs a specific task, such as statement spreading, an expert system.

For the majority of bankers involved in commercial lending, however, expert systems are an attempt to capture the thought processes of experts and make that expertise available to other users. Commercial loan expert systems technology makes an evaluation and aids in the credit decision-making process of loan officers. It is with that definition in mind that the phrase expert systems is used throughout this article. Slow start. The technology of expert systems is just one of six subdisciplines of artificial intelligence, although it is the most well known and has the most application to banking. The others are speech recognition, natural language processing, robotics, computer vision, and software development.) With expert systems, users have access to the knowledge of experts in a particular subject area through the use of computer terminals. In a bank, that would make the knowledge of the most experienced members of an institution readily available to less experienced loan officers to aid in their credit decision-making process.

When talk of expert systems started several years ago, there was hope that they would be installed in banks across the country right away. However, the once bright outlook for expert systems and for the AI industry in general has failed to materialize. In fact, according to The Wall Street Journal, the industry has been plagued by slow growth and big cutbacks in the past five years. Analysts once predicted this would be a $4 billion industry by 1990; however, general estimates of the market today are closer to $600 million.

According to the Journal, companies failed for a number of reasons, including a poor understanding of the potential markets and overpriced products. Lack of knowledge. There are several reasons bankers are wary of expert systems technology. Robert Long, president of Thinking Technology Associates, Phoenix, says one reason is that "a lot of people just don't know anything about expert systems. " In his book Artificial Intelligence for Commercial Lenders, published by Robert Morris Associates, Long gives four reasons commercial lending managements are not embracing expert systems technology. Long says these reasons surfaced from interviews with commercial banking executives and with Al vendors who called on bankers:

(1) Present technology is already burdensome, and the bank's data processing and systems group is too busy to undertake new projects.

(2) Lending is an art, and no computer will ever be intelligent enough to be of significant service.

(3) Banks are not developers of systems. Bankers will use an expert system when someone brings it in and convinces them that it is beneficial. …