Process design should consider alternative flowsheet topologies and alternative equipment sizes. For instance, the process design of a distillation sequence should consider different sequences (direct, indirect, others), different overhead compositions, recycle flowrates, reflux ratios, etc. Each design will have its own energy consumption. Moreover, each design will have its own potential for integration with the rest of the process. To keep track of all alternatives, the process design engineer usually relies on simulation. Different process topologies and equipment sizes are simulated, costed and compared. The wider the scope of the system is the more difficult it becomes for the engineer to keep track of all solutions and the search becomes "ad hoc". In this situation, process integration helps. Process integration methods often guarantee, or nearly guarantee, optimal structures for a significant part of the system. In other words, process integration plays an important part in process design. However, there have been barriers to overcome in the introduction of process integration to the process design environment. Perhaps the most important such barrier has been the problem of "data extraction". The task of "data extraction" presented a barrier because it required expert understanding of process integration techniques, particularly Pinch Technology. Inexpertly extracted data often led in practice to situations where optimum solutions were missed. Moreover, the task of "data extraction" was time consuming, even for the expert. In 1997, Linnhoff March launched software to perform the skilled task of "data extraction", automatically. This navel and unique software is based on an expert system which acts as an intelligent interface between process simulation and process integration. In 1998, Linnhoff March complemented this with more-direct links with the main commercial process simulators. The overall package analyses any simulated process at the press of a button, automatically, to identify the potential for improvement. There are many benefits. Design time is saved. The impact of process changes on overall cost is quickly analysed for any number of options. The power to do this is made accessible to a larger number of less-experienced users. All in all, process designs will be designed close to optimum, by less experienced users, in less time. In this paper, we outline the essential principles for extracting process integration data from process simulations, as implemented in Linnhoff March's expert system software. Practical examples from industrial projects demonstrate these principles to produce rapid and reliable results to reduce process energy levels and capital investment.