In order to innovate and respond quickly to new requirements, employees frequently supplement their information systems. This particularly applies to the context of business intelligence (BI) because many users supplement their BI platforms with individually tinkered spreadsheets. Unfortunately, these supplements bear numerous threats such as limited report reuse across all potential users. To address this gap, we establish a design science project. First, we qualitatively explore impediments to diffusion of reports and impediments to innovative use. Second, upon our findings and extant literature, we derive meta-requirements for BI platforms that foster diffusion of reports and innovative use. Third, we develop and discuss principles for how to design a BI platform that would meet the identified meta-requirements. The resulting design principles emphasize (1) permanent user sandboxes to improve innovative use and (2) hybrid recommendation agents based on user interaction, collaborative-filtering, and users' social influence to improve diffusion of reports.