One of our main goals is improving data economy readiness in regional companies in Norway. As part of our initiative, a collaboration was initiated with a local power company on inflow prediction. Hydrology is a complex topic and there is interest in investigating how accurately domain knowledge in hydrology and machine learning can be used to improve reporting of production capacity to power exchange markets. Our contribution is an approach on how researchers can help companies get started with machine learning. It borrows practices from agile software development methods.