Search

The explainability of artificial intelligence (AI) is one of the central challenges for the wider use of the new technology in many industries and applications. The more powerful and efficient the algorithms of AI work, the less it is usually comprehensible to users. While there is widespread agreement on the basic requirement of explainability for AI applications, the design of an adequate AI explanation is rarely defined. This contribution presents basic concepts of explainability as well as current approaches to explanations for AI. It describes which methods are fundamentally suitable for considering an explanation to be complete and how it must be designed in order to be assessed as interpretable for AI.

What is special about the German-speaking education system is that there are two types of higher education institutions. While classical universities do academic teaching, so-called universities of applied sciences (UAS) focus more on practical training for the labour market. This chapter presents the experiences of two professors with practice-integrated teaching at two UAS that ensure employability to a large extent. The common setting of the four courses presented is that students work on real business cases, practical challenges, and real fields of innovation. Methodologically, design think and SCRUM are used. It is described for these cases how the design of the course contributes to students developing innovative and digital approaches for real companies. The chapter concludes with a compilation of good practices that emerged from experiences. These can serve as input for other teachers to design their own practice-integrated courses.

The processes of an investment bank are considered to be particularly knowledge-intensive, because analysts need to extract or generate relevant knowledge from a variety of data. With increasing digitization, modern data science and business intelligence techniques are available to support or partially automate these activities. This study presents concrete use cases for front office processes of an investment bank as how knowledge management techniques can be used. For example, the article describes how expert systems can be used in the due diligence review or how fuzzy logic systems help in deciding whether to buy or sell securities. The article is based on 1079 texts (e.g. documented cases and articles) and serves researchers as well as practitioners as an application overview of data science techniques in the example area of knowledge-intensive banking processes.

Data-driven business processes management (BPM) is regarded as a central future trend because automation often makes huge amounts of data (big data) available for the optimisation and control of workflows. Software manufacturers also use this trend and call their solutions big data applications, even if some features are reminiscent of traditional data management approaches. This chapter derives from the basic definitions of big data including 13 central requirements that a big data BPM solution must meet in order to be described as such. One hundred twenty-one process management solutions are evaluated on the basis of these to determine whether they are real big data applications. As a result, less than 5% of all solutions analysed meet all requirements.

Data science as the interdisciplinary collection of methods and techniques to support businesses is becoming more and more popular. This article begins with definitions and shows how systematically competitive advantages can be built up on the basis of digital data. Essential sources and types of data-driven knowledge are introduced. Then a classification of approaches of data science concepts is explained. A distinction is made between Business Analytics and Business Intelligence as different levels of analytical skills. The paper goes into depth with these concepts and presents concrete techniques, algorithms, and application scenarios. Thus, the contribution introduces State of the Art approaches to analysis, control, monitoring but also to advanced approaches such as prediction, simulation, and optimization.

A key objective of strategy management is ensuring the implementation of strategies in the day-to-day operational business processes of enterprises. For this purpose, researchers and practitioners developed an enormous body of methods for strategy management. However, the tremendous variety of strategy management methods makes it difficult to choose the most appropriate methods for a given challenge. In an effort to provide guidance for practitioners and researchers, this paper provides an integrated dataset of the 50 most widespread strategy design and implementation methods and links these methods to a large body of 100 case studies. Based on this dataset, commonly utilized methods, common challenges, typical objectives, and specifics of industries are examined.

Process orientation is often seen as a central starting point for optimizing business performance. This is especially true for supply chain workflows because modeling, understood as specifying activities, data, resources, and control flows, is usually the first step in selecting and implementing digital approaches such as smart factories, virtual agents, and autonomous systems. Many different process notations are known (e.g., detailed vs. rough, flexible vs. standardized, etc.), and at the same time, fundamentally different SCM process types make various assumptions about suitable modeling (e.g., integration of value chain partners, standardization). This chapter presents an association model that suggests which notation fits best to which SCM type. The chapter discusses the benefits of specific notations for classical as well as modern challenges of SCM to help in the selection of notations. The resulting specifications serve as the basis for the holistic digitization of supply chains.

The picking process in warehouses is an important step in global supply chains because it causes significant costs and has a high impact on customer satisfaction. Especially manual systems may cause high costs and can generate many picking errors such as type and quantity failures. Using the example of consignment at an international manufacturer located in Switzerland, this contribution provides in-depth fundamentals on picking processes. The picking system, typical workflows of manual picking procedures and theoretical considerations of optimal picking quality are given. Using a quality cost as well as a process cost analysis and a cause and effect analysis, the picking process is investigated in this scenario. It turns out that the optimal picking quality is surpassed in the case, which requires a systematic process re-design in order to cut costs. Therefore exemplary technical, organizational and soft best practices are shortly presented. In the scenario a business and technical case of the introduction an innovative weighting system is described.