This thesis applies computational intelligence to the field of actuarial (insurance) science. In particular, this thesis deals with life insurance where mortality modelling is important. Actuaries use ancient models (mortality laws) from the nineteenth century, for example Gompertz' and Makeham's law of mortality, to calculate lump sums and premiums for life insurances. In this thesis, the process of modelling is revived by eliciting new knowledge of the parameters in Gompertz' mortality law. This new knowledge leads to an easy applicable law in the theory of heterogeneous populations. Therefore, separate groups of policyholders get risk premiums (and corresponding policies) adapted to their risk profile. A step further: This knowledge leads in this thesis to a new explaining model for heterogeneous populations (by using a vitality parameter) where even the physiological properties of the policyholder can be summarised. This knowledge results in a more than ever individualised mortality prognosis (for example, BMW drivers who smoke cigars with have a hearing aid). In this thesis, the WWW sets in because of the need of - in the future - selling insurance policies via PC. To do this, more intelligence has to be inserted in the premium calculation, to make it sufficiently fast. This is done by introducing so-called germ functions. With these germ function formulas for life insurance can be automatically derived by an autonomous softbot-clerk.