02.03.2013 Views

Downloadable - About University

Downloadable - About University

Downloadable - About University

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

442 Alternative decision-support systems<br />

One development of the system is to automate the majority of the<br />

underwriting process. Details from a life proposal form can be entered<br />

into the system at a branch office by a clerk with no underwriting<br />

experience. Evidence is automatically sent for where required, with<br />

details of complete medical reports, questionnaires, etc. also entered<br />

into the system by a clerk.<br />

The automated processing, which gives ‘instant’ decisions on 85% of<br />

applications (as opposed to a previous baseline of 65% instant decisions<br />

by clerical staff who consulted a simplified underwriting ‘screening’<br />

manual) leads to increased speed and consistency of application processing<br />

for substantially more customers. In addition, since insurance<br />

brokers typically send off several proposals for an individual customer,<br />

the first positive decision back to a customer has a business advantage.<br />

Another calculation revealed that, since the senior underwriter was<br />

much less comprehensive in this request for medical evidence (at a cost<br />

of roughly $75) than the deliberately conservative ‘screening’ manual, a<br />

saving of many hundreds of thousands of dollars could be achieved by<br />

letting the senior underwriter in the expert system decide when further<br />

medical evidence was required.<br />

Some companies are now processing 90% of their new life applications<br />

through expert systems. Extant systems are now able to perform electronic<br />

data interchange (EDI) with third-party information providers<br />

such as screening laboratories. The biggest benefit is seen to be time<br />

saving in the turnaround of life proposals.<br />

However, another argument is that expert systems that enhance speed<br />

of turnaround time should not be deployed without careful consideration<br />

– since customer relationships need to be built for long-term<br />

retention. Such relationships have, at least in the past, been built on<br />

face-to-face meetings of field agents and potential customers – where<br />

policy alternatives are discussed and confidence in subsequent recommendations<br />

is established. 27<br />

The expert systems built by Nippon Life 28 and by the Swiss Reinsurance<br />

Co., 29 among others, have attempted to produce fully automated<br />

systems. In these cases, the system builders have attempted to automate<br />

the entire life underwriting procedure, by encoding their extensive<br />

underwriting manuals into knowledge bases.<br />

The extent to which this has proved feasible is still an open question,<br />

for the law of diminishing returns applies at a certain stage, where it<br />

becomes less and less economic to encode the rules required to deal<br />

with the last few percent of all cases (which was left, in one system, 30<br />

for expert underwriters to deal with). However, none of the systems

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!