Decision-Making Business Software May Be What the Doctor Ordered
As a surgeon in Detroit several years ago, F.N. (Buzz) Burt wondered why certain doctors asked the right questions during patient consultations while other doctors seemed unable to gather information needed to make a diagnosis.
Burt’s interest in patient interviewing techniques evolved into a study of the human decision-making process, and Burt eventually abandoned his successful medical practice to found Cogensys, a San Diego-based company that is developing software systems that emulate human decision-making processes.
But Burt isn’t marketing the Cogensys system, which he describes as a form of artificial intelligence, to doctors.
Instead, Cogensys has focused on the lucrative market it expects to develop as financial institutions turn to computers to help them make better decisions on which consumers are credit-worthy and which clients are good insurance risks.
Potential Called Huge
The company, with headquarters in La Jolla, has identified 22,000 banks, saving and loans, mortgage bankers and insurance companies as potential customers for its software programs. The software “captures and emulates” the decision-making logic that humans use to solve problems, Burt said.
In the past, expert systems were “rule based,” meaning that the systems were unable to deal with new problems. Cogensys claims that its patented software can rapidly adapt itself as, for example, interest rates change or the dollar’s value drops.
Financial institutions don’t expect knowledge-based systems such as those being developed by Cogensys to “replace every decision-maker in the world,” according to Ed Nichols, chief of information services at San Diego-based Home Federal Savings & Loan. However, the systems will free up loan officers and underwriters who now spend too much time on relatively mundane decisions, Nichols said.
The driving forces behind Cogensys’ marketing program are the increasing number of bad loans being made by S&Ls; and banks and a wave of underwriting problems that threatens to swamp some insurance companies.
Bad Judgment Blamed
“All we have to do is point to the Sheshunoff Lending Report (a bank rating service) which blamed bad loan judgment for the majority of bad loans,” said Cogensys President and Chief Executive Donald W. Fuller. “Loan losses and charge-offs are increasing every quarter.”
Nearly a quarter of the nation’s life and casualty-property companies face short- or long-term financial difficulties, according to the results of a recent survey by the Washington, D.C.-based National Assn. of Insurance Commissioners.
Financial institutions increasingly are turning to artificial intelligence--the computer science that strives to duplicate human intelligence--to reduce the bottom line consequences generated by bad decision-making.
For example, Citibank last Wednesday agreed to pay $2 million to New Haven, Conn.-based Cognitive Systems Inc., which is developing artificial intelligence software for use by financial institutions. Cognitive Systems hopes to produce “knowledge-based, natural language software systems that allow people to communicate with computers in everyday language,” according to a company spokesman.
Revenues for the still-evolving artificial intelligence industry are expected to total $4 billion by 1990, according to industry analysts. Cogensys describes its “expert systems” software as one type of artificial intelligence.
Little Research So Far
However, financial institutions have been slow to fund research into expert systems, Fuller acknowledged. During 1986, banking, finance and insurance companies pumped just $18 million into expert systems, Fuller said.
That reluctance stems from the fact that past AI products have not been very practical, Fuller said.
Cogensys’ patented software is practical because it incorporates “the thought processes that experts use to make decisions,” according to Burt. “As an expert trains the software, the system captures the inexpressible elements of human judgment, as well as the explicit ones.”
“The system is really like a brand new baby when it starts up,” according to one insurance industry executive who is familiar with Cogensys’ products. “It learns as it interacts with the expert. It gets even stronger as it looks at more situations, which sets it apart from any other system I’ve seen.”
Fuller said Cogensys has sold about a dozen of its software-based systems in the United States and Europe but declined to project what the company’s sales might total this year.
Cogensys has created an underwriting decision-making system for Alexander Hamilton Life Insurance Co., a Detroit-based subsidiary of Chicago-based Household International Inc., that began screening insurance applications two weeks ago, according to Alexander Hamilton Vice President Mike Wilson.
“My first reaction was that it sounded like something out of ‘Star Wars,” ’ said Wilson, who added that more than a third of the nation’s insurance companies now are working on artificial intelligence applications that could one day make underwriting decisions.
Policy From a Machine?
Alexander Hamilton is using the Cogensys software system to screen simpler applications, but the company hopes to one day create an “expert system” that “takes people’s money and gives them a policy right there at an ATM-like machine,” Wilson said. “I think that’s going to be conceivable with some of our more simplistic and lower-cost policies.”
Cogensys also is developing a system that will let Phoenix-based Valley National Bank provide instant credit decisions to automobile buyers. Dealers will key in information, and the Cogensys software will search credit bureau reports and other databases to determine if a consumer is credit-worthy.
That kind of instantaneous decision-making capability will be welcomed by car dealers who log a hefty percentage of their sales on weekends when banks and other loan sources are closed, according to George Mullinax, a senior vice president who oversees consumer lending at Valley National Bank.
