Experiments at synchrotron beam lines are commonly steered by humans. Since beam time is expensive, the human experts have to work without interruption for days to make decisions about the measurements. Still, in complex situations, the decisions, the expert makes, are sub-optimal. A computer can keep perfect track of the conducted measurements and suggest optimal future measurements. With this technology beam line scientists can now run a software and spend time on other things, return later and find the optimally explored model of a material or object. To achieve this, the algorithm takes advantage of statistical and machine learning methods coupled with advanced function optimization.