AI in New Product Development : Connecting Data & Unlocking Knowledge

Abstract

Companies engaged in complex product development are increasingly striving to automate their R&D processes and transform them through the integration of AI. This transformation promises substantial acceleration of product development cycles, higher product quality, improved compatibility between mechanical, E/E and software components, and enhanced capabilities for design space exploration. While numerous approaches for embedding AI into engineering already exist, most organizations have not yet established a comprehensive management framework or achieved large-scale deployment of this technology. This white paper introduces a framework for scalable AI applications in engineering, designed to address the unique challenges of product development environments. Considering the specific boundary conditions in engineering, such as fragmented tools and data landscapes, heterogeneous data formats, stringent governance requirements, complex tool interoperability, and high interdependencies between engineering domains along the V-model, the framework identifies key dimensions that must be addressed to ensure scalable and sustainable AI integration. Following terminology from Systems Engineering, the framework distinguishes between vertical and horizontal AI use cases, depending on their level of domain specificity and cross-domain applicability. Based on an extensive literature review covering AI use cases across all domains of the V-model, the paper highlights that most existing AI applications currently exhibit a high vertical maturity but limited horizontal integration. In other words, they are typically designed around specific tools and data sources within isolated domains, with insufficient focus on cross-domain networking and knowledge sharing. This pattern mirrors the current state of industrial AI adoption, which is often constrained by tool and data fragmentation as well as organizational silos. Consequently, the paper argues that AI transformation must be driven by top management, ensuring a balanced portfolio of vertical and horizontal use cases and promoting integration across domains to unlock system-wide benefits. A further key insight of the study is the emerging role of multi-agent systems in engineering, which enable higher levels of automation and coordination between AI-driven tasks. This white paper illustrates how such systems can be applied in future engineering environments and analyzes their impact across four dimensions, namely processes, data, people, and technology. Finally, a roadmap is presented that outlines the path toward scalable AI deployment in product development. In conclusion, the paper recommends a strategic and iterative approach to AI transformation: selecting and developing use cases in alignment with the proposed framework, progressively interconnecting them into multi-agent systems, and ensuring governance and scalability from the outset. To achieve lasting success, organizations must also make deliberate choices regarding the right tools and technologies and understand the correct sequence for generating and structuring the required engineering artifacts. Providing contextual information across different system levels is essential to enable consistent interpretation and automated reasoning. Furthermore, the long-term integrability of initially developed use cases must be safeguarded to ensure they can evolve into interconnected multi-agent ecosystems rather than remain isolated solutions. Finally, human checkpoints embedded into agent-driven workflows play a pivotal role in maintaining oversight, trust, and accountability, ensuring that automation augments rather than replaces engineering expertise. Datei-Upload durch TIB