Data science projects are pivotal in generating value through the effective use of data. Despite substantial investments in data science initiatives, many organizations face challenges in realizing the full potential of these projects. This study unfolds how organizations develop data analytics capabilities. Using a qualitative research approach, we conducted 38 in-depth interviews with informants from various sectors engaged in data science projects. Our findings highlight key capabilities that data science teams must develop, including collaboration with do-main experts, AI ambidexterity, agile project management, and privacy competencies. In addition, we show that as data science teams performing in the organizational context, in their day-to-day tasks experience challenges: lack of data understanding, data bottlenecks, and data sustainability. Such challenges trigger organizations to develop three capabilities: democratization, enhancing data-driven culture and balancing risks and innovation.
