Pharmaceutical development is based on complex relationships between active pharmaceutical ingredients (APIs) and excipients material properties, process parameters and targeted product performance which are difficult to model and predict. With advancement of the knowledge base in the field, availability of large datasets and machine learning, digital platforms have been created providing advanced algorithms to simplify and accelerate formulation development. In this study, the applicability of BASF Virtual Assistant ZoomLab® [1], as a tool for development of directly compressible formulations was investigated, using caffeine and rivaroxaban as model drugs. Comparative evaluation of the experimentally obtained and built-in data has been performed using the API and Powder Blend Tabletability and Processability applications followed by Formulation Wizard guided formulation selection. Based on API characteristics evaluation, both caffeine and rivaroxaban presented poor processability, indicating that direct compression into tablets would be challenging. However, Formulation Wizard identified a number of prospective formulations with good processability. Powder blends, selected based on predicted favorable flowability or tabletability, exhibited good flowability, powder density, and compression-related properties. The obtained tablets exhibited good mechanical properties (hardness and friability), as well as fast disintegration. ZoomLab® built-in algorithms, as well as the API and excipient databases represent useful preformulation tool and provide insights which could guide formulation development. However, the excipients database is limited to the selected BASF products. Relevant outputs should be critically assessed and further optimized based on experimental observations leading to hybrid approach which would accelerate product development while minimizing time and resources consumption.