Integrating Assessment Methods in the Development of ML-based Business Models for Manufacturing

Abstract

The use of machine learning promises great potential along the entire value chain of manufacturing companies. Many companies have already recognized the resulting opportunities for increasing enterprise value and are developing their machine learning applications for the production environment. However, despite these efforts, many of the solutions developed fail in the market. Especially small- and medium-sized enterprises have difficulties developing suitable business models for their technical applications. These difficulties arise because companies do not evaluate their business projects sufficiently during the development phases. As a result, unpromising projects are not recognized until late in the development process and thus cause high sunk costs. This paper presents an approach for integrating assessment methods into developing machine learning- driven business models for production. Due to the diametric evolution of information availability and uncertainty during the business model development process, various methods and tools can be used for the assessment depending on the current phase. For this purpose, existing assessment methods are evaluated and contrasted regarding their suitability concerning machine learning-based business models for production. Afterwards, three approaches for the different planning phases of business model development (strategic, tactical, operational) are presented in this paper.