In today’s increasingly digitalized world, companies have to deal with ever enormous data floods, which eventually lead to a data overload undermining decision quality of their employees. The petroleum industry in particular faces the problem of an aging workforce with more experienced people retiring than fresh engineers stepping in. This means that within the next years fewer engineers have to cope with increasing volumes of data and information in a more and more complex operational environment and the permanent loss of accumulated knowledge of the retiring experts. The work suggests that a solution lies in the focus of knowledge management’s efforts on the level of information technology (IT). Data, information and knowledge form a pyramid where the amount of data generally decreases towards the top, whereas the structure and logic increases. Today's IT systems are stuck at the information level of the pyramid. Knowledge can be automatically generated from data and information and universally reapplied to new situations. Pushing information technology towards a knowledge technology (KT) with highly adaptable and learning network structures could bring a long required software innovation to companies. The work emphasizes that with intelligent systems, companies learn how to learn faster, thus secure a sustainable competitive advantage. As a practical example of the implementation of KT, the Adaptive Advisory System is introduced. It is a decision support system for oil and gas operations that makes use of artificial intelligence to infer knowledge from data and information, store and adapt to new challenges and therefore steadily increases the quality of results along with its usage. It is designed to complement oil company’s assets teams for highly complex problems with limited expert availability, but with a great flexibility of application. In addition to user-system-interfaces, it comprises from a Data, an Information and a Knowledge Layer, with the latter being the centerpiece of the Advisor. The Knowledge Layer consists of the three different sections: the Event Detector identifies real-time data deviations and alerts in case of emergencies. Combining these events, the Problem Classifier assesses the situation and infers the most likely problems. The Decision Supporter calculates the utilities of actions and recommends the best solutions to the user. In a final evaluation, the decision projection is compared to the actual results some time after the action to further improve the results. Event Detector, Problem Classifier and Decision Supporter are based on Bayesian networks, an instrument of artificial intelligence, that is widely and successfully used for knowledge representation and reasoning under uncertainty. Bayesian networks are defined as acyclic directed graphs with nodes (random variables) that are interconnected by edges (conditional probabilities). The work describes the principles, structures as well as the operating modes of the Adaptive Advisory System and illustrates these by examples from the area of well production monitoring. It proposes several ideas for a user interface in order to meet the requirements of a transparent and flexible Advisory System. Finally the focus is put on the human factor, since the best system is worthless without being accepted by the user.