Diese umfassende Fallstudie taucht in das Gebiet des Wissensmanagements (KM) ein und dessen Anwendung bei Wiewald, einem Unternehmen, das sich auf die Planung und Gestaltung von Kompressorsystemen spezialisiert hat. Die Studie präsentiert eine umfassende Untersuchung verschiedener Dimensionen des KM und bietet strategische Empfehlungen, die auf Wiewald’s Digitalisierungsstrategie zugeschnitten sind.
Die Studie hebt die Bedeutung des KM im digitalen Zeitalter hervor, in dem die transformative Kraft der digitalen Revolution einen Paradigmenwechsel in der Fertigungsbranche herbeigeführt hat. Das Aufkommen des Internet-of-Things und der Zustrom von Massendaten haben den Übergang von Big-Data zu Smart-Data erforderlich gemacht, welche handlungsorientierten und aussagekräftigen Informationen repräsentiert. Es wird jedoch darauf hingewiesen, dass Organisationen das volle Potenzial von Smart Data möglicherweise noch nicht vollständig ausgeschöpft haben.
Die Fallstudie zieht Erkenntnisse aus empirischer Forschung mit malaysischen Fertigungsunternehmen heran und hebt die entscheidende Rolle der Integration des KM in Geschäftsprozessen hervor. Vier Dimensionen des KM werden identifiziert: Wissensschaffung und -erwerb, Wissensaustausch und -transfer, Wissensspeicherung und -ab- frage sowie Wissensanwendung. Diese Dimensionen sind miteinander verbunden und betonen die Bedeutung der Nutzung organisatorischen Wissens für effektive Innovationen.
Die akademische Literatur zum KM wird durch eine bibliometrische Analyse erforscht, um Lücken zu identifizieren und aktuelle Trends sowie aufkommende Techno- logien zu integrieren. Die Studie betont die Notwendigkeit, die Kluft zwischen praktischen Leitlinien für die Wissensanwendung in der Industrie und dem Stand des Wissensmanagements in der akademischen Forschung zu überbrücken.
In Zusammenarbeit mit Wiewald konzentriert sich die Studie auf die Entwicklung eines Kompressorsystem-Konfigurators, der dem Cheftechniker bei der Planung und Gestaltung neuer Systeme helfen soll. Durch Machbarkeitstests unter Verwendung historischer Daten wurden in der Vergangenheit Entscheidungsbaummodelle entwickelt, die basierend auf grundlegenden Anforderungen wahrscheinliche Entscheidungswege bestimmen können. Dieser Konfigurator zielt darauf ab, Vorschläge für gültige Systeme zu liefern und schnelle Reaktionen sowie hochwertige Entscheidungsfindung zu ermöglichen.
Für die Implementierung des Konfigurators wird ein Ansatz namens „Actionable Cognitive Twin“ (ACT) vorgeschlagen, der die Transparenz von Informationen, die Entscheidungsunterstützung und die Agilität verbessern soll, während er gleichzeitig die Beteiligung der Mitarbeiter an der Gestaltung von Geschäftsprozessen und die Förderung von Innovationen anregt. Die Implementierung von ACT umfasst die systematische Verwaltung von Wissen, die Integration von Ontologien und Wissensgraphen sowie den Einsatz datengetriebener Techniken wie Natural Language Processing und neuronalen Netzwerken.
1 Background
Digital innovations and technologies, especially artificial intelligence, present a unique opportunity to fundamentally change how organizations work: improving knowledge access, facilitating more extensive product and service development, minimizing bureaucratic obstacles and enabling the emergence of new business domains. Put simply, artificial intelligence plays a key role in facilitating a digital revolution towards a knowledge-driven economy.
Small- and medium-sized enterprises (SMEs) play a vital role in the German economy, accounting for 99.4 percent of the market and driving national competence development [1, 2]. Fully capitalizing on this opportunity for digital transformation is crucial given their significant influence, which is why the German government’s strategic initiative for Industry 4.0 [4], aims to leverage and maximize SMEs’ transformative potential. It is reasonable to assume that by embracing Industry 4.0, SMEs can ensure long-term competitiveness in the evolving business landscape, positioning themselves for sustained success.
Anzeige
Unfortunately [3], “Let There Be Change”, sheds light on the limited AI implementation among German companies. The study reveals that a mere 12 percent of these companies consider their AI maturity level sufficient to improve their processes and organizations. In contrast, a substantial 63 percent remain in the experimental phase of adopting AI at an enterprise level, indicating their challenges in leveraging AI effectively.
The empirical study by [5] with 120 senior executives from Italian manufacturing firms highlights the crucial role of knowledge management practices (KMPs) in achieving AI maturity and leveraging its potential. The findings show that AI deployment positively influences KMPs, stimulating their higher levels while enhancing supply chain resilience (SCR) and manufacturing firm performance (MFP). This emphasizes the need for an effective digitalization strategy that prioritizes KMPs as a mediating mechanism. By leveraging KMPs to transform data into valuable knowledge, organizations can enhance decision-making and improve MFP and SCR outcomes.
The KMI project (“Künstliche Menschlich Intelligent” – federal grant no. 02L 19C500) champions sustainable AI; prioritizing worker benefits over profit-driven approaches like Industry 4.0. By augmenting worker capabilities, reducing menial tasks, and ensuring usability, sustainable AI aims to preserve employability, fostering a more worker-centred “humane” approach within the Industry 5.0 context, where worker benefits are valued alongside company profits [4].
Building upon the insights of [5], we explore the potential applicability of their findings in a real-life SME. Wiewald, a specialist in configuring, installing, and maintaining compressor systems, serves as our focus for this investigation. As an active application partner in the KMI project, our research revolves around enabling the sustainable integration of AI within SMEs, drawing from our own insights and experiences gained through our work with Wiewald, which have shed light on additional challenges commonly faced by SMEs in the project.
Anzeige
One such challenge is an insufficient ‘data-mindedness’, referring to the unaware- ness among SMEs about the opportunity to generate additional value from their data for various reasons. In the case of Wiewald, this could be attributed to unstructured data or the absence of digitization. Another challenge is the presumed absence of a structured and integrated process to externalize and document employee knowledge, particularly when tacit expert knowledge is required for projects involving new innovations such as AI.
Concerning these two issues, our experience at Wiewald and a new understanding concerning the importance of knowledge management practices for digital transformation, we believe that companies like Wiewald, need to rely on knowledge and its exploitation, to sustain a long-term advantage. While extensive knowledge management practices may not be essential for day-to-day operations, SMEs must proactively prepare themselves with a digitalization strategy to remain relevant in the evolving landscape of the German economy’s digital revolution. This strategy ensures their long- term success in an emerging knowledge-based economy, where the ability to leverage knowledge and adapt to digital advancements becomes a crucial competitive advantage.
2 Hypothesis and Research Question
We anticipate that an improved availability and accessibility of data and information would concurrently improve the potential for value generation and MFP of SMEs, thereby creating the necessary foundation for digital transformation strategies. Hence, we hypothesize that a viable solution must include:
A means for systematic knowledge management to capture explicit knowledge and preserve tacit expert knowledge within the company, explicitly integrating it into a business process model.
A holistic measure for innovation management that encourages active involvement from workers, especially when structuring or modelling business data and business processes, ideally facilitating worker-driven initiatives for innovation management.
By reconsidering our approach regarding a successful digital transformation strategy in SMEs, an argument concerning the merit and strategic importance of knowledge management practices for the sustainable implementation of AI should be raised and evaluated, leading to the following research question:
“To what extent can the effective and decisive application of knowledge raise an SME’s AI maturity level and determine the success of its digital transformation?”
Evaluating this research question involves assessing the impact of effective knowledge application on AI maturity and its correlation with the success of digital transformation efforts in SMEs. The examination will consider factors such as the percentage of explicitly captured business process knowledge compared to processes lacking explicit detailed information, as well as tracking the number of adjustments and innovations introduced by workers since the inception of the digitalization solution, along with their impact on MFP.
This research question delves into the relationship between knowledge management practices (KMPs), AI maturity, and the success of digital transformations in SMEs. It seeks to demonstrate the crucial role of KMPs in elevating an SME’s AI maturity level and achieving a successful digital transformation.
To address this research question, the case-study will provide a concise and comprehensive overview of knowledge management, exploring its various dimensions, inter- connections, and their impact on driving process and product innovation. By doing so, it will examine how effective knowledge management enhances AI maturity within organizations and facilitates the development of impactful digital transformation strategies.
3 Theory: Global Perspective – Comparing Industry Experience with Structure of Academic Research
The digital revolution, driven by advancements in technology such as the Internet of Things, has brought about a transformative shift in the manufacturing domain [7]. It has enabled seamless communication and autonomous collaboration among machines, leading to a massive influx of data commonly known as big data. However, it’s important to note that big data is not synonymous with smart data, which represents actionable and meaningful information. Despite the prolonged existence of the digital revolution, there remains a possibility that organizations have not fully transitioned towards harnessing the potential of smart data [10].
As organizations are encouraged to adapt to the emerging knowledge-driven economy, investigating the transformation from big data to smart data becomes essential. This investigation involves examining the differences and similarities between contemporary theoretical perspectives and practical experiences, providing insights into the evolving role of knowledge management in the digital era. By exploring this interplay, we can comprehensively understand how the digital revolution has influenced knowledge processes within manufacturing companies, thus establishing the theoretical framework to effectively address our research question.
In their empirical work, [6] conducted a questionnaire involving 162 Malaysian manufacturing firms, highlighting the significance of integrating knowledge management (KM) into business processes. They concluded that KM plays a crucial role in driving process- and product innovation, leading to innovative value creation. Their findings provide valuable insights and guidelines for effective KMPs.
The study defines four dimensions of KM (shown in Fig. 1), which serve as the foundation for our case study. These dimensions examine how organizational knowledge is: created and acquired; shared and transferred; stored and retrieved; and applied. Their findings emphasize the importance of interrelationships among KM dimensions for their overall effectiveness, increasing the firm’s potential to leverage the inherent value of their data for effective innovations [6].
Fig. 1
Overview of the 4 knowledge management dimensions, their interrelationships and influence on technological innovation, according to [6]
×
[7] contributes to the prior industry experience by providing their academic perspective. They conduct a bibliometric analysis of 90 relevant articles and use cluster analysis to explore the intellectual structure of academic literature related to knowledge management in Industry 4.0. Their analysis results in the identification of six clusters of relevant keywords, which serve as the basis for their systematic literature review. Their research allows them to identify gaps in the existing academic literature and integrate current trends and emerging technologies, ideas, and concepts. This enables them to provide insights into the future directions that the knowledge management literature may take.
Therefore, by comparing the practical guidelines for applying knowledge management in industry with the state of knowledge management in academic research, it is possible to integrate emerging concepts and ideas related to KM. This integration allows for the identification of parameters that contribute to the development of a robust digitalization strategy applicable to SMEs in general, and specifically to Wiewald.
3.1 Knowledge Acquisition & -Creation
The academic sphere places significant emphasis on ‘knowledge creation’, with approximately 50% of the papers in the dataset addressing this topic [7]. The creation of knowledge involves the generation of new knowledge within the company or its acquisition from external sources. By adopting this broader definition of knowledge acquisition, its link to innovation is reinforced, as the creation of knowledge significantly enhances innovative performance [6]. Consequently, when internal and external knowledge sources are effectively combined, the acquisition of knowledge exerts a positive influence on innovation.
For this case study, ‘knowledge acquisition’ involves discovering, locating, creating, or capturing new information, including both intangible tacit knowledge and tangible explicit knowledge from diverse sources such as employees, data, and external entities [6, 7]. [7] primarily focusses on ‘knowledge capture,’ highlighting its positive relationship with ‘knowledge sharing’ and ‘-transfer’, with guidelines promoting an effective knowledge-sharing culture that is impactful to decision-making and thereby relevant to the ‘knowledge application’ dimension [6].
‘Knowledge creation’ trends overlap with ‘-transferability’ trends, such as smart and digital environments. This results in two sub-streams of academia; process modelling and condition monitoring, which prioritize cyber-physical- and semantic technologies like digital twins, semantic webs, and semantic web rule languages to enhance ‘knowledge acquisition’ capability [7].
3.2 Knowledge Sharing
‘Knowledge sharing’, also known as ‘knowledge accessibility,’ allows employees to share and access both tacit- and explicit knowledge within and outside the organization [7]. The presence of a culture of ‘data-mindedness’ acts as a moderating variable, enhancing the speed of information flow, thereby emphasizing the need for an integrated transparent knowledge-sharing behaviour [6].
This transparency creates a ‘proximity effect’, enabling increased access to information and faster knowledge dissemination during day-to-day operations [6]. Contem porary views highlight the paramount importance of ‘knowledge sharing’, evident in its prevalence across academic research and its influence on various trends [7].
Although the transfer of knowledge is often regarded as the most significant knowledge management dimension [6], its direct impact on strategic objectives, such as achieving higher AI-maturity, may be debatable.
3.3 Knowledge Storage
‘Knowledge storage’, or ‘knowledge documentation,’ is vital in the knowledge management process, involving capturing, refining, structuring, integrating, and storing information [6]. This concept resonates with academic topics such as big data, digital transformation, and cyber-physical systems that are heavily influenced by the digital revolution [7].
The concept of ‘organizational memory’ involves capturing tacit knowledge and enhancing its accessibility, thereby preventing the loss of valuable information [8]. This, in turn, fosters ‘organizational agility’ by enabling organizations to assimilate up-to- date knowledge and respond swiftly to new information. By combining these practices with the ‘proximity effect’ and fostering a culture of ‘data-mindedness’, knowledge storage practices accelerate the transmission and dissemination of knowledge, thereby driving innovative behaviour [6].
It is reasonable to assume that within such a dynamic environment, shaped by effective knowledge storage, that organizational knowledge can be leveraged to make well- informed decisions for long-term sustainable growth and an immediate competitive edge.
3.4 Knowledge Application
In the context of the global paradigm shift driven by the digital revolution, the importance of ‘knowledge application’, also known as ‘knowledge responsiveness’, becomes increasingly pronounced [6]. It represents a strategic asset for knowledge-driven organizations, facilitating a realignment of goals towards long-term strategic value creation. This realignment extends beyond the scope of business intelligence; which focuses on immediate data utilization, and enters the realm of data science. Through data science, organizations can generate profound insights that drive innovative advancements in business processes and products. Beyond supporting organizational goals, such strategic utilization of knowledge promotes collaboration and innovation, leading to the development of new products and processes. In fact, research has shown that effective knowledge application facilitates the translation of organizational expertise into innovative advancements, contributing to overall organizational innovation [6].
This transformative process towards advanced data-driven decision-making, includes the concept of ‘responsiveness’ that encompasses both ‘agility’ in reacting to changes in information and the ‘quality of response’; which empowers organizations to take strategic actions and unlock new possibilities of higher quality, agility and effectiveness [6].
While the Malaysian industry questionnaire recognizes the strategic significance of ‘knowledge application’, ranking it as the second most important dimension in relation to innovation [6], the academic literature on KM reveals that it is the least investigated dimension, with only 16% of the papers in the dataset referencing knowledge application [7]. This discrepancy underscores the need to address this potential research gap and explore how the effective application of knowledge can bridge the gap between information advantages and actionable decisions.
Drawing insights from the investigative results of [5] and [7], we align our under- standing with the current state of knowledge management practices (KMPs) in the field. These results emphasize the vital role of ‘knowledge application’ in achieving impactful outcomes in the digital era [3, 5]. It becomes evident that SMEs need to prioritize a digitalization strategy that effectively applies their knowledge resources to create a competitive advantage and foster innovation. By doing so, they can thrive in the rapidly evolving digital landscape. These findings underscore the significance of strategic knowledge management and its relevance in driving success for SMEs in the digital age.
4 Implication and Conclusive Recommendation for Wiewald
In partnership with Wiewald, we have been working on a compressor system ‘configurator’, that is supposed to assist their chief technician during the planning and design of new compressor systems. With early feasibility tests using historic data that was available to us at the time, we were able to design simple decision tree models that were able to determine most likely decision paths, when given basic requirements about desired compressor capacities. This initial work has sharpened our vision for the digital solution that suits Wiewald’s requirements.
As the chief technician elicits the requirements from their customer, he enters the relevant conversation data into his electronic quotation form. The configurator uses this input to perform calculations in a discrete backend service. Once the basic datapoints are determined, the configurator should be able to provide suggestions for valid compressor systems.
The benefits of this configurator will be demonstrated by its rapid responsiveness and the quality of decisions it facilitates. With this prototype system, the customer should be able to gain an accurate understanding of the anticipated system price and its complexity before the session of the quotation process has concluded. Further adjustments to the requirements can be explored during the very same conversational session. This should reduce a two-week process with multiple tedious back and forth requirement-elicitation sessions (as shown in Fig. 2) into a single transparent session in which requirements and potential solutions can be communicated effectively and dynamically. Furthermore, the session can be concluded with the signing of a handover confirmation form. Requirements elicitation is central to their business process landscape; its successful completion determines and triggers all necessary business processes.
Fig. 2
Approximation of the current proposal creation process at Wiewald
×
As of recent, we have received a rich and qualitative bounty of new data that we aim to utilize for the creation of a new prototype of superior sophistication. Our approach is inspired by the designs of the ‘Actionable Cognitive Twin’ (ACT) proposed by [9]. ACT establishes a solid foundation that enhances information transparency, decision support, and agility, while also ensuring that workers actively participate in the shaping of business processes, thereby driving innovative initiatives. This integrated approach aligns with the goal of effectively utilizing digitalized organizational domain knowledge and fostering a culture of data-mindedness and innovation:
It captures both explicit and tacit expert knowledge within Wiewald, providing a means for systematic knowledge management and its explicit integration into their business process model; leveraging ontologies, knowledge graphs, and embeddings to externalize implicit information, thereby fostering a comprehensive understanding of the business processes. This integration of knowledge promotes effective ‘knowledge capture’ and ‘-preservation’, ensuring that organizational expertise is embedded within the company’s business operations.
Furthermore, ACT serves as a holistic measure for innovation management, encouraging active involvement from workers, especially during the structuring and modeling of business data and processes, and facilitating worker-driven initiatives for innovation. This empowers employees to contribute their valuable insights and ideas, creating an environment that fosters collaboration and drives innovation within the organization. Beyond promoting worker empowerment, the implementation of ACT enables the adoption of AI-solutions to further enhance innovative organizational capabilities
To implement ACT, we would employ a data-driven approach, that leverages natural language processing (NLP) techniques such as entity recognition, relationship extraction, and semantic parsing. Through these techniques, we would extract relevant information from collected data and utilize neural network-based ontology learning methods to automatically learn ontologies from structured data and textual resources. These ontologies capture complex relationships and hierarchical structures, providing a foundation for knowledge representation. Furthermore, the ontological learning process would include the creation of an ontological embedding, which represents the distilled information from the learned ontology. This embedding serves as valuable information for an informed neural network (INN), guiding its understanding and reasoning capabilities.
The INN, equipped with the ontological embedding, is responsible for the capture and analysis of organizational data, leveraging the knowledge captured in the embedding to interpret and make sense of the incoming data. A knowledge graph serves as an instance of the learned ontology and is dynamically populated and updated by the INN. By populating the knowledge graph with relevant data from real-time sources (like the requirements elicitation form), the INN would create an interconnected representation of the acquired information.
The knowledge graph, shaped by the INN, provides a structured and comprehensive view of the information captured from the interactive data streams, representing an integrated view of the entire organization’s knowledge. Therefore, it should enable systematic knowledge management, facilitate decision-making, and foster innovation within the organization. Furthermore, continuous refinement and validation of the domain knowledge ontology, knowledge-graphs, and ontological embeddings, in collaboration with domain experts, will be needed to ensure their accuracy, relevance, and coverage.
By implementing the configurator as an instance of ACT; an integrated approach, combining ontological learning, ontological embedding, and INN-driven knowledge graph population, we aim to unlock new possibilities for effective knowledge representation, reasoning, and information retrieval for the task of compressor system configuration at Wiewald.
Open Access Dieses Kapitel wird unter der Creative Commons Namensnennung 4.0 International Lizenz (http://creativecommons.org/licenses/by/4.0/deed.de) veröffentlicht, welche die Nutzung, Vervielfältigung, Bearbeitung, Verbreitung und Wiedergabe in jeglichem Medium und Format erlaubt, sofern Sie den/die ursprünglichen Autor(en) und die Quelle ordnungsgemäß nennen, einen Link zur Creative Commons Lizenz beifügen und angeben, ob Änderungen vorgenommen wurden.
Die in diesem Kapitel enthaltenen Bilder und sonstiges Drittmaterial unterliegen ebenfalls der genannten Creative Commons Lizenz, sofern sich aus der Abbildungslegende nichts anderes ergibt. Sofern das betreffende Material nicht unter der genannten Creative Commons Lizenz steht und die betreffende Handlung nicht nach gesetzlichen Vorschriften erlaubt ist, ist für die oben aufgeführten Weiterverwendungen des Materials die Einwilligung des jeweiligen Rechteinhabers einzuholen.
