Based on an analysis of building processes to date, this paper proposes two ways to make Building Information Modelling (BIM) and with it complex building processes less expensive and more reliable. First, existing experience with building processes that include building-integrated solar systems is analysed, as they are one example for buildings with high energy-saving goals. Based on the analysis, the authors propose to include a new property set and functions in machine code in the next version of the Industry Foundation Classes (IFC). For the machine-code functions, several formats are proposed and the advantages and disadvantages of the proposed IFC extensions are discussed. For simple conventional building components, the existing IFC version 4 seems appropriate. Buildings with innovations that add complexity to the building process can profit most by the savings generated by exchanging machine-code functions. In building processes to date, much information is not shared because of proprietary matters. Functions in machine code can provide the necessary confidentiality so that more capabilities can be exchanged between the stakeholders. Models based on machine-code functions can be updated much faster than detailed models with many parameters. With machine-code functions, advantage can be taken of the greater accuracy of detailed models, as long as they are fast and easy to use. The modularity of the functions stimulates competition between alternative approaches, which makes simulation models less expensive. Finally, exchanging a closed, validated simulation model in machine code introduces a much lower risk of error than exporting and importing complex simulation models. (C) 2017 Elsevier B.V. All rights reserved.