Founded on the pre-master project Industrial Internet: Sensor Applications and Measurements in a Workshop Setting (A1) as a starting point for thinking about early stage product development in the industrial internet, this thesis tries to bridge the gap between research and hype on one side and a practical approach to prototyping. Wayfaring was used as the main method for gaining insights on how an advanced industrial machine was used and how data may be incorporated in feedback loops or repositories. Machine learning was used on a dataset of 165 instances of machine uses to predict familiarity with the machine. The learner was not able to significantly beat always choosing the majority class. Still, insights into how to prototype the data gathering and handling may be used by others to improve the interplay between man and machine. Data was collected from six sources to monitor usage of a laser cutter. A coding scheme were developed to quantify each user session. This prototype driven-approach allowed for prototyping, testing, and learning in two dimensions with data and without data. Insights from probing and need finding with users produced a rich description of what users want from a machine and how they are using it. Further attention should be on using adaptive interfaces and feedback relevant to certain users. This area was only touched-upon and will likely be a promising venue for further research in an industrial internet context.