The paper considers digital tutoring platforms as commercial products in the shadow education sector. In particular, it focuses on relatively new platforms which purportedly act as tutors themselves: web-based learning software providing "intelligent tutoring systems.' The question is not how effective and beneficial these platforms are for learning success. Following science and technology studies, the focus is on product design and its role as an actor in a network of relations. Autoethnographic interaction with a paradigmatic platform is used to show a variety of inscriptions and translations. The classifications, standardizations and quantitative calculations that emerge here are analyzed as practices for generating data, which in turn stabilize the platform as a specific ontological configuration. The data give the impression of being personal. On closer examination, however, they are the product of manifold relations between different actors in the network.