The paper presents novel techniques to process planning problem specifications, expressed in a declarative description language, which enables the description of planning problems with incomplete knowledge. The outcome is improved performance and scalability of conformant planners. The paper proposes two transformations of a planning problem specification, aimed at reducing the size of the initial belief state and the number of actions to be dealt with. The two transformations have been implemented in a static analyzer and in a companion heuristic search conformant planner (CpA+). The performance of the resulting system is compared with other state-of-the-art conformant planners.