The interactive simulation of flexible surface-like components is becoming increasingly important in early phases of automotive product development, in particular for digital assembly planning and virtual prototyping. Typical examples include trim parts as well as interior and exterior panels, which may undergo large rigid-body motions combined with significant elastic deformations during assembly. While detailed nonlinear shell formulations provide a physically consistent description of such behavior, their computational cost prevents their direct use in interactive simulation environments. In this work, we present an approach that enables real-time interaction with flexible surface-like parts by combining high-fidelity shell-based modeling with an efficient reduced representation. The underlying formulation is based on geometrically exact shell theory and serves as the reference model for model generation, validation, and offline analysis [1]. For interactive simulation, only a reduced shell model is derived by static condensation (Guyan reduction) [2], resulting in a condensed linearized formulation. Although the reduced model is linearized, it is applicable in a practically relevant deformation regime that exceeds the standard applicability limits of purely geometrically linear models that would admit only tiny deformations w.r.t. the reference shape. This allows for interactive manipulation of flexible parts while preserving essential mechanical characteristics. The reduced formulation is not intended to replace the underlying nonlinear shell model, but rather to act as a computational surrogate that is specifically tailored to interactive use. The method is embedded into an industrial simulation workflow and applied to representative assembly scenarios in which flexible panels are grasped and deformed to reach their final mounted position. In addition to computational aspects, practical issues such as model preparation, robustness during interactive manipulation, and integration into existing simulation environments are addressed. The presented work demonstrates how advanced mathematical models can be transformed into practically usable simulation tools and illustrates the role of model reduction as an enabling technology for mathematics-driven innovation in industrial applications.