| dc.description.abstract | In these hard economic times, the Kenyan Insurance Broking Industry is currently
undergoing a difficult phase. Good service delivery, retention and expansion of their clientele
is the key to their basic survival. This includes prompt selection of reliable insurers to place
the clients' businesses with.
Research was done on how the big 5 insurance brokers evaluate the insurance companies that
they use to place covers for their clients. Only 5 companies out of the registered 187
insurance brokers in Kenya were chosen as they had the biggest market share and were
amongst the few highly computerized insurance brokers. Of these 5, 2 are multinationals and
the other 3 are local companies. Currently in Kenya there is no automated methodology in
use to evaluate and select insurers. This process is done manually and only by the top
underwriting or technical managers - "The Experts".
The purpose of this study was to develop a framework of how expert systems can be used in
the underwriting industry in Kenya. It gives details of how the methodology, system design
and system testing of an Expert System is carried out. To demonstrate how this framework
can be applied a limited functionality of an insurer evaluation expert system was developed
using fuzzy logic. Fuzzy logic methodology was used in the expert system to accommodate
the vagueness of the natural language and the intuitiveness of human knowledge. In addition
it will also be shown that Fuzzy logic is a simple way of handling rules in an expert system.
Results show that four main factors are used in evaluating Insurers. The Financial Strength,
the Management Quality, the Reinsurance and the Claims Attitude of the Insurer.
Kenya is not abreast with Expert system technology not because of lack of qualified IT
professionals but more so because of the lack of awareness that such technology exists. This
study also aimed at bringing awareness to both the Insurance Industry at large and the
Kenyan IT group of professionals that this technology can be used successfully on a local
level | en |
