This policy brief presents a succinct account of the potentials and limitations of data-driven policy learning emerging as an innovative paradigm able to harness the power of data and predictive analytics to enhance ex-ante policy design – such as the optimal individuals to support or the optimal treatment to provide – across diverse socio-economic domains. By integrating advanced technologies such as machine learning and causal inference techniques, policy makers can assess the potential impacts of different policy options on a wide range of societal, economic, and environmental sectors. This shift towards evidence-based ex-ante policy design has the potential to revolutionize decision-making processes. The FOSSR project envisions the creation of a Policy Learning Platform (PLP) that will bridge the gap between recent theoretical developments in policy learning and their practical application in real-world policies. Using this platform, policy-makers can identify potential risks and trade-offs before policy implementation, thereby refining policy execution and basing decisions on empirical evidence, which promotes transparency and accountability.