The need for adaptiveness of business applications is on the rise with continued increase in business dynamics. Code-centric techniques show unacceptable responsiveness in this dynamic context as business applications are subjected to changes along multiple dimensions that continue to evolve simultaneously. Recent literature suggests the use of product line architectures to increase adaptiveness by capturing commonality and variability to suitably configure the application. Use of model driven techniques for developing business applications is argued as a preferable option because platform independent specification can be retargeted to technology platform of choice through a code generation process. Business applications can be visualized to vary along five dimensions, namely, Functionality (F), Business process (P), Design decisions (D), Architecture (A) and Technology platform (T). Use of models is largely limited to F and P dimensions in commonly used model-driven development techniques thus limiting the benefits of product line concept to these two dimensions. We argue this is not sufficient to achieve the desired adaptiveness, and it is critical to extend the product line concept to D, A and T dimensions also. To address adaptation needs of business applications, this paper presents a model-driven generative approach that further builds on the ideas of separation of concerns, variability management and feature modeling. Early experience and lessons learnt are discussed, and future work outlined.