The design of an industrial robot involves a number of conflicting objectives in general. A parallel kinematic machine (PKM) is known to achieve high precision and heavy load capacity but with the sacrifice of large workspace and high dexterity. Therefore, existing PKMs are mostly dedicated to a specific task with relatively poor adaptability to task variations. In this paper, the concept of an adjustable platform is proposed to enhance the adaptability of a PKM for various tasks. It is demonstrated that the adjustment of the dimensions of a base platform or end-effector platform has a significant impact on the performance of a PKM including its workspace and overall stiffness distribution. Both offline and online adjustment modes are presented. The offline adjustment brings a new dimension for reconfigurability and thus increases the versatility of a parallel robot for different tasks. The online adjustment turns a parallel robot into a redundant PKM, and therefore its overall performance against task requirements can be improved. A planar PKM and a Stewart robot with an adjustable platform are used as case studies in order to demonstrate the advantages of the proposed concept.