Managing Artificial Intelligence

Inhaltsverzeichnis

1. Frontmatter

2. 1. Introduction to Managing Artificial Intelligence

   Nils Urbach, Daniel Feulner, Philipp Dilger

   Abstract

   This chapter sets the stage for understanding AI and its management. It explores the rise of AI, defines its core concepts, and traces its historical milestones. These foundational insights are followed by a discussion of the essential themes in AI management, structured around the key areas covered in this book: establishing a technical foundation, identifying and designing AI use cases, developing AI strategies and governance frameworks, translating AI strategies into functional solutions, and scaling them effectively. Together, these elements provide a comprehensive framework to navigate the complexities and opportunities of AI in a business context.

3. AI Foundations

   1. Frontmatter

   2. 2. Technological Foundations of AI

      Nils Urbach, Daniel Feulner, Tobias Guggenberger

      Abstract

      This chapter explores the core technical components that make AI possible, including data quality, storage architectures, computational power, and tools for AI development and deployment. It examines key learning methods—supervised, unsupervised, and reinforcement learning—alongside foundational algorithms such as regression, classification, and clustering, which drive AI-driven insights and automation. The goal is to provide a structured introduction to AI’s technological foundations, enabling readers to understand its core mechanisms and make informed decisions about its application.

   3. 3. Foundations of Neural Networks

      Nils Urbach, Daniel Feulner, Tobias Guggenberger

      Abstract

      This chapter examines the key components of neural networks, including input, hidden, and output layers, as well as the role of weights and activation functions in the learning process. Particular emphasis is placed on backpropagation as a core training technique. These concepts are explored in a structured manner, illustrated through the example of identifying handwritten digits, demonstrating how neural networks process and learn from data. The goal is to provide a clear and comprehensive introduction to the principles and mechanics of neural networks.

   4. 4. Introduction to Generative Artificial Intelligence

      Nils Urbach, Daniel Feulner, Simon Feulner, Tobias Guggenberger, Valentin Mayer

      Abstract

      In this chapter, we examine the evolution and impact of generative AI, starting with how ChatGPT became the catalyst for generative AI. We then move beyond ChatGPT to discuss foundational concepts and models such as large language models (LLMs), diffusion models, and the emerging field of agentic AI. Understanding these technologies is essential in AI management, as they significantly influence strategic decision-making, risk management, innovation opportunities, and operational efficiencies within organizations.

   5. 5. Evaluating and Optimizing Artificial Intelligence Models

      Nils Urbach, Daniel Feulner, Tobias Guggenberger, Annalena Schmid

      Abstract

      In this chapter, the initial focus is on how AI models fit into a broader hierarchy, distinguishing them from foundational mathematics and algorithms. This is followed by a discussion of the key challenges in AI model development and the central role of metrics in addressing them. Next, various metrics are presented in detail, along with hyperparameter optimization methods aimed at improving model performance. The final section highlights the characteristics of a well-trained AI model, emphasizing strategies to avoid both underfitting and overfitting for robust, reliable results.

4. AI Ideation

   1. Frontmatter

   2. 6. Application Potentials of Artificial Intelligence Technologies

      Nils Urbach, Daniel Feulner, Simon Feulner

      Abstract

      In this chapter, we explore how AI contributes to the transformation of business operations by enhancing decision-making, increasing efficiency, and enabling new forms of value creation. Using practical examples, we illustrate how AI technologies, such as perception, recognition, decision-making, and prediction, are applied to streamline and enhance complex workflows. Moreover, we explore AI’s integrative role in innovative, cross-technology applications, particularly in conjunction with blockchain, the Internet of Things (IoT), and quantum computing.

   3. 7. Identifying, Designing, and Evaluating AI Use Cases

      Nils Urbach, Daniel Feulner, Simon Feulner, Dominik Protschky

      Abstract

      In this chapter, we outline a structured approach that guides organizations through three key stages of AI use case development. First, we focus on the identification of promising use cases by combining organizational needs with AI capabilities. Next, we address the design of AI-based services using the AI Service Canvas, which helps structure data requirements, business potential, and organizational integration. Finally, we present the evaluation of AI use cases through the effect-path model—a tool for systematically tracing how data and AI capabilities generate business value and competitive advantage.

5. AI Strategizing

   1. Frontmatter

   2. 8. AI Strategizing and Readiness

      Nils Urbach, Daniel Feulner, Valentin Mayer, Simon Meierhöfer

      Abstract

      This chapter begins by introducing the concept of AI strategizing and presents a taxonomy that helps organizations conceptualize the design space of an AI strategy along four dimensions: scope, scale, speed, and source. Building on this, this chapter defines AI readiness as a multidimensional construct that captures an organization’s preparedness to successfully adopt and scale AI. The final section analyzes key factors that influence AI readiness—including strategic alignment, resource availability, and cultural adaptability—and discusses how these factors can be assessed and developed to support effective AI use in practice.

   3. 9. Governance and Management of AI

      Nils Urbach, Daniel Feulner, Simon Feulner, Moritz Schüll, Valentin Mayer

      Abstract

      This chapter explores governance mechanisms that support the safe and effective use of AI as well as presents a method for transforming existing governance approaches in response to the rise of AI. Building on this foundation, the chapter then introduces the AI application management (AIAMA) model as a comprehensive framework for managing the AI lifecycle in alignment with governance principles and organizational goals. It outlines core management dimensions of AI, including technical, process-related, and user-centered aspects, and explains how these are coordinated through integration management. Finally, the chapter discusses the design of organizational structures for organizing AI efforts, including centers of excellence, cross-functional, virtual, and matrix teams.

6. AI Design and Development

   1. Frontmatter

   2. 10. Techno-Economic Decisions of AI

      Nils Urbach, Daniel Feulner, Simon Feulner, Valentin Mayer

      Abstract

      This chapter explores the key factors influencing such decision-making in the context of AI integration. It begins by introducing the “machine learning decision space,” which highlights strategic considerations in balancing performance, robustness, and interpretability across the AI lifecycle. It then examines infrastructure models—such as edge computing, on-premises hosting, and cloud computing—that support data management, model training, and deployment. The role of AI service platforms is also discussed, focusing on how cloud-based, pre-built solutions can accelerate implementation and reduce complexity. Finally, the chapter introduces the concept of MLOps, which bridges the gap between development and production to ensure reliable and scalable AI operations. Together, these elements form a comprehensive framework for navigating the techno-economic landscape of AI adoption.

   3. 11. Designing Human–AI Interactions

      Nils Urbach, Daniel Feulner, Valentin Mayer

      Abstract

      This chapter explores the dimensions of human–AI interactions, identifies distinct interaction types, and discusses influencing factors such as transparency, personalization, anthropomorphism, and user expectations. It then examines the transformative impact of AI on the workforce, highlighting the necessity of collaboration to manage job displacement and innovation effectively. Furthermore, the chapter presents strategies for optimizing human–AI collaboration, including managing automation bias and ensuring appropriate reliance on AI systems. Finally, it addresses effective task delegation approaches, emphasizing conditions for successful collaboration and analysing AI-led delegation through a principal–agent perspective, ultimately aiming to support practitioners in designing sustainable and ethically responsible human–AI interactions.

7. AI Operations at Scale

   1. Frontmatter

   2. 12. AI Monitoring and Change Management

      Nils Urbach, Daniel Feulner, Annalena Schmid, Dominik Protschky

      Abstract

      This chapter addresses three core pillars of managing AI at scale. First, we discuss key performance indicators (KPIs) that link AI initiatives to strategic goals and operational performance. Second, we outline an iterative approach to machine learning monitoring, detailing how to observe, evaluate, and improve model behavior in dynamic environments. Finally, we introduce a structured framework for AI-related change management, offering practical tools to navigate employee concerns, foster acceptance, and anchor AI sustainably within the organization.

   3. 13. Ethical, Legal, and Social Implications of AI

      Nils Urbach, Daniel Feulner, Valentin Mayer

      Abstract

      This chapter begins with a discussion on ethical foundations, societal roles, and normative frameworks relevant to the responsible development of AI systems. It then explores artificial moral agency, examining whether and how AI systems can be designed to make morally informed decisions. Building on this, the focus shifts to design considerations for transparent and fair AI systems, guided by ethical norms and technical standards. The chapter also outlines the conditions necessary for trustworthy AI, emphasizing alignment with societal values. It concludes by addressing legal implications, including data protection, intellectual property, and new regulatory obligations under frameworks like the EU AI Act.

8. 14. Correction to: Designing Human–AI Interactions

   Nils Urbach, Daniel Feulner, Valentin Mayer

   Dieses Kapitel geht auf die Szenarien eines komplementären Potenzials zwischen Mensch und KI ein und zeigt auf, wie sie effektiv zusammenarbeiten können. Es stellt verschiedene Arten der Integration für menschliche und künstliche Intelligenz dar und bietet eine visuelle Darstellung durch angepasste Zahlen. Der Text betont, wie wichtig es ist, diese Wechselwirkungen für eine bessere Entscheidungsfindung zu verstehen. Darüber hinaus bietet es Quellen und Referenzen für weitere Explorationen, was es zu einer wertvollen Ressource für Fachleute macht, die ihr Wissen über die Zusammenarbeit zwischen Mensch und KI erweitern wollen.

   KI-Generiert

   Diese Zusammenfassung des Fachinhalts wurde mit Hilfe von KI generiert.

   Originaltext anzeigen