The Role of Artificial Intelligence in the Knowledge Organization of Companies
Introduction
The increased importance of human resources, knowledge, and intellectual capital figure is a challenge for organizations in a competitive business environment. The primary challenge is the maximum utilization of the intellectual capital by requiring professional organization of knowledge. The question is how to capture and utilize the knowledge and experience stored in human brain? How can accumulated knowledge be used? This challenge indicated the development of IT systems that are intended to provide access to documentable knowledge and information. Real business success requires development of integrated systems that can answer all the issues of knowledge management (availability, sharing, development, storage, utilization, and completing strategic objectives). This integrated system is not just a standard enterprisewide development and use of IT solutions, but a new way of utilizing human resources and intellectual capital. It has moved the organization toward the creation of a new knowledge-based organization and building knowledge management system (KMS).
Theoretical Background
“There is nothing new under the sun” might be an applicable statement since knowledge management is not a recent phenomenon. The KMS is a new system approach and way of thinking that refines the previous views and management tools.
The development and historical stages of KMS are mainly based on the study of Anklam (2005). In his opinion, the evolutionary stages of knowledge management have been completed and currently the development is in phase six.
The first phase is characterized by a focus on technology anchored on generating knowledge (Nonaka and Takeuchi 1995) with the end in view of knowledge utilization. Knowledge is viewed as a product and information is referred to as a resource.
The main characteristic of the second phase is a conscious handling and recognition of differences in knowledge-based, experiential and problem-solving knowledge. The problematic question is how to encourage the human resources to share their knowledge (Poór 2010).
In the third phase, knowledge is perceived as a network (Chena et al. 2014), where the cooperating organizational partners integrate business models, complex structures, and innovation system groups.
The fourth phase following the classification of Anklam focuses on knowledge as a capital factor and is trying to quantify it, while the fifth phase focuses on the relation between corporate competitiveness and innovation. It is clear that developed economies in the fourth and the fifth phase are looking for the importance of the value of human resources. The current highest level of development is phase six (not yet defined by experts, but seemingly feasible), which due to the effects of Industry 4.0 and digitalization highlights the inevitability of artificial intelligence (AI) (Barkovics 2016). Each step of the KMS system involves those new solutions, which find a solution for problems with the help of digital technology that could not be handled by early IT systems.
The question is how will development proceed. One possible option is the return to the first phase (dominance of IT system building) where IT will gain importance in a new form. With cloud-based solutions, AI will take over the dominance, providing solutions to the questions and problems formulated earlier. There is talk about cyclically repeating demands, but always at a higher level. The earlier phases of development are emerging again, but in the form of KMS supported by AI. The value of human knowledge in this context comes to different light and the possibilities of system building with the support of AI also require a new way of thinking. The other option is to continue the progress and reach phase seven of development, where the effects and expectations can be hardly predictable.
The mentioned phase evolved sequentially as a natural consequence of development of technology and thinking. Nowadays, one can see a positive experience and initiatives in companies representing countries with different culture and level of development, which responding to the expectations of Industry 4.0 can act as best practice for others (Bencsik 2013).
In the following sections, the logic where building KMS is treated as a self-returning cyclical development process will be shown. With the help of this approach, an illustration of how KMS steps can be combined with the possibilities provided by digitalization, that is, how to integrate the solutions of AI into the process of corporate knowledge management, will be shown. An example on how the logic supported by AI can be implemented will be provided. Finally, a summary of possible solutions and methods that paired with the tools applying KMS from a table will be provided.
Knowledge Management System Model
The definition of KM defines an activity chain that presents the knowledge management as a progressing, evolving, and cyclical process (Davenport 1996; Duhon 1998; Gholami et al. 2013). This logic focuses on developing and managing the organizational knowledge base. One of the best-known models was developed by Probst et al. (2006). It consists of eight components that can be divided into external and internal cycles. The individual components are interconnected and form a system based on a logical principle. Following the logic of the model the content requirements of the elements are introduced.
External Cycle
Knowledge Objectives
The knowledge management objectives determine what kind and level of skills and knowledge to develop and use to reach organizational objectives for the next strategic period. Different objectives should be selected at different levels: normative, strategic, and operative.
AI helps with setting objectives, especially in case of “what if” type of questions. Systems supporting strategic decisions bring rules and logic to process data sets. It helps the automation of certain decision-making situations and is also suitable to prepare forecasts (one possible tool is “SAP Digital Boardroom” digital decision support on three screens).
The next phases of the KMS model are always connected to the phase before. This is a start of the internal cycle, including the following six steps.
Internal Cycle
Identification of Knowledge
This component of the model involves an overview of the existing organizational skills and knowledge capital. Before starting a new project, it is necessary to map the existing knowledge. Nowadays, information technology makes it possible to store data exactly in a form it was created. To make the knowledge more transparent, it can be helpful to use tools that can provide an overview about the availability of necessary knowledge (Hanako 2016).
As corporate knowledge is expanding rapidly, a teachable AI system can be established. This system would enable access to customized corporate knowledge. After submitting the required information and based on the answers provided for questions, the system will automatically be able to build the organizational knowledge base, which will automatically be extended with new information (Az öntanító AI 2017). A single, self-learning system will prevent the loss of information, will remember and put each of the knowledge items into the appropriate place. It will keep the organizational knowledge up to date and the holders of knowledge can achieve strategic objectives or the system can be customized to search for information from external sources (SAP HANA solution).
Knowledge Acquisition
Useful information is acquired through formal and informal channels in the company and will make finding knowledged individuals essential for the organization. These channels can be used internally and externally (Davenport and Prusak 2001).
Biometrics allow people and machines to develop even more natural interactions, such as touch, image recognition, speaking, and body language. Currently it is primarily used in market research, but it also supports the selection of suitable competencies and human qualities (selecting the suitable workforce, choosing an expert). The role of AI is to process the huge amount of statistical data based on specified parameters (Kiss 2017).
Knowledge Development (Knowledge Creation)
The purpose of knowledge development is to ensure that employees of the company will generate the knowledge they need, including ideas, models, skills products, processes, and so on by choosing new or traditional methods of learning (training, e-learning, distance learning, school education) (Lovaszy 2017).
This purpose can be served well through follow-up that supports the application of AI solutions. The machine-based learning can have various forms (Hasznics and Nuridsány 2006). In the case of neural networks, a pattern-based learning is characteristic. The goal is to gain the appropriate knowledge from large amount of data and to change the behavior of a system through it (Watson’s solutions).
Knowledge Sharing
Knowledge sharing is a critical phase of the knowledge management cycle and can be realized directly or indirectly. No matter how it is realized, people, organization, and technology are the key factors.
An important area related to AI is the expanding knowledge-sharing solution: a machine intellect is learning from other AI via the Internet. It is an easier way for machines to be taught the new assembly line process with this method than programming them one by one. The cooperative robots (cobots) were designed to work with people as their direct associates, preventing them from repetitive, burdensome, and dangerous operations. The cobots are connected to AI cloud applications, so they can be taught new work processes without programming (Tölgyes 2016; Kulaklia and Mahonyb 2014; Bíró 2017).
Preservation and Fixing of Knowledge
The main objective of knowledge preservation is to make the knowledge accumulated by the company accessible and ready to be utilized for many years. The explicit knowledge can be easily stored in an accessible form. A bigger problem for a company is retaining knowledge, which is hidden in the minds of employees. In the case of knowledge that is nonconvertible into an explicit form, organizational memory plays an important role.
This phase includes fixing of knowledge, systematization, storage, and update. These knowledge-based systems store the explicit form of knowledge of different problem areas or store expertise knowledge to handle a specific problem area. The inference systems are the catalysts of knowledge-based systems. For example, the central feature of case-based reasoning (CBR) is to solve the present problems based on the solutions of similar past problems (real situations and interpretations of them) (Ojala 2016; Stone et al. 2016). The system is capable of self-learning and is able to learn from experience if users provide feedback that solution offered was successful or not.
Use of Knowledge
The utilization of knowledge should ensure that knowledge is used productively and contributes to the improvement of the company performance (Fehér 2003).
Properly used knowledge is an integral part of the company’s activity. The user-friendly nature of technology that makes knowledge accessible can help employees in the utilization of knowledge (i.e., using text-based AI for process efficiency), speed of problem solving, and enhancement on the quality of customer service (OTP Bank, Contact Center systems). Aside from intelligent search function, which can be applied in complex sentences or exact terms, the system is capable of proposing replies for e-mails and chat conversations and visualizing data sets or intelligence escalation, which can make certain processes more efficient (Hoeschl and Barcellos 2006; Hlács 2016).
Evaluation/Measurement of Knowledge
The last element of the model—reviewing the purpose of the process and knowledge—is usually not given the appropriate importance. The assessment and control of knowledge increase the visibility of the changes in organizational knowledge.
When evaluating the competitiveness of the organization, one must not forget that indicators measured in the traditional way will only partially show the real value of the company. Characteristics associated with human resources fall into the invisible category of the company’s balance sheet. It would not be fair to dismiss them when evaluating the performance of the company, since the value of the organization is strongly associated with its employees and their ideas. The value of this invisible capital without humans is worthless or simply does not exist (Birzniece 2011; Neururer 2015).
Integrating AI—The Case-Based Reasoning
CBR is modeling the human thinking, which is a process of solving new problems based on the solutions of similar past problems and tasks by searching for and adopting documents from the past. It can be applied for the KMS in the following form. During the knowledge identification phase, the characteristic features describing similarities (codes) help to find the bearers of knowledge in the organization and through knowledge acquisition (if a new employee or expert is needed) find the staff to be recruited. It can identify the missing or outdated knowledge during the phase of knowledge development. During knowledge sharing, the solutions developed earlier are uploaded, and a predefined structure is applied and made available for use in the database later. A knowledge store is created, which makes available the previously tried and tested solutions, utilizing knowledge in the current case (Kiss 2017). The conditions to use the method efficiently are the following:
- Appropriate coding of cases
- Quick access to right solutions, tasks (qualities, competencies, etc.)
- Determining the degree of similarity
- Ensuring simple adaptation
- Constant update of the case-base
The CBR process is very simple. Individual cases are stored—together with the concrete solution of the cases—to create a case-base. When solving a new task/problem, the most similar case is selected from the case-base and the solution of the past case is adapted to the new one. Finally, the new results are added to the case-base (Badinszky 2008).
Representation of Cases
A problem to be solved during the development of an expert system based on CBR is the representation of cases.
A graph is used to describe the characteristics of the problem or task. The complexity of the graph depends on the complexity of the processes in the company and is prepared considering the characteristic features of the organization. A dialog (question-graph) will help the user to progress toward the solution of the problem, the code.
Identifying Similarity
The success of identifying similarity will fundamentally influence the success of the entire process and can therefore be considered as a key problem in the process. To solve this problem, an artificial neural network can be applied.
Artificial Neural Networks
Teaching is a prerequisite for the use of neural networks. Teaching patterns are used during the teaching process. These patterns include the primary similarity indexes and the final similarity indexes as well. This pattern is generated automatically using the case-base.
The Connection Between AI and the Building Blocks of KMS
Table 1.1 summarizes the steps of KMS, the processes (see above) required to implement them, and some examples of useful tools. Parallel with the digitalization areas, solutions are also presented supporting the most up-to-date solutions utilizing the possibilities provided by AI. The support provided by AI enables the attainment of a high level of knowledge organization and management that it can compete with human skills or exceed its boundaries in quality, quantity, and access to data and information. The AI solutions applied in each step and the basics of these solutions serve as examples that may change, develop, and multiply or in some cases disappear. It provides a framework for developers of KM systems and can contribute to the decision on the method to be applied (Mi az OCR Technológia 2017).
Table 1.1 Toolkits and relationship between KMS and AI
|
Digitalization areas |
Artificial intelligence (AI) solutions (examples) |
|
|
Probst’s building blocks of KMS and their processes |
Methodological background, management toolkit (examples) |
|
|
1. |
Networking, decision models, brainstorming |
OCR (optical character recognition), text production, deep learning, identity management, digital boardroom, probability networks, and so on |
|
2. |
Benchmarking, knowledge map, competency map, databases, data warehouses, document management systems |
Machine learning (self-teaching system), content-based systematic search, intelligent response system, keyword recognition and self-reflection, big data solutions, cloud-based service |
|
3. |
Corporate records, learning, monitoring competitors, market information, Peer Assist, lessons learned, after-action review (AAR) |
Text analysis, biometrics, chatbot, content-based semantic search, textmining, competitor intelligence (CI), translators, speech recognition (hidden Markov models), HR Hacking |
|
4. |
Learning, e-learning, training, brainstorming, talent programs, career plans, development programs, language learning, development of professional communities, Lifelong learning, blended learning |
Follow-up systems, learner-teacher systems, chatbot, multi-agent intelligent tutoring system, MOOC (massive open online course), advisory artificial intelligence |
|
5. |
Knowledge sharing, on-the-job, video sharing, video conferencing, Wiki, World Café, storytelling, blogs, documentation systems, discussions, forums, brainstorming, ROCK (retention of critical knowledge), SharePoint, Yammer |
Virtualization, speech recognition systems, text analysis, Google Docs, SharePoint, Social Learning, MyNet, GrapeVINE, intelligent escalation system, DeepCoder, case-based reasoning (CBR), artificial neural network, genetic algorithm, language translators, speech recognition (hidden Markov models), deep learning |
|
6. |
Registers, work instructions, repositories, catalogs, printed and electronic forms, archive training and teaching materials |
OCR, expert systems, DeepCoder, knowledge-based system, CBR, interactive search engines, personal knowledge management, and content developing software |
|
7. |
AAR, retrospect, lessons learned, informal discussions, teamwork, big data |
IoT (Internet of things), content management softwares, speech recognition softwares, intelligent response systems, cobots, virtualization, speech recognition (hidden Markov models) |
|
8. |
Analytical systems, measurement and management models, measurement decomposition structures |
IoT, content management softwares, neural networks |
Conclusion
The potential of AI provides easily accessible tools for KMS. The development in this field is extremely fast and relies on the creativity of executives and how they can adapt to the processes that have been developed. A number of user-friendly and innovative solutions are available to the public and selecting the appropriate expertise is essential. This expertise will lead to the creation of tools not only for KMS but for other areas of organizational knowledge management as well. The self-learning feature of AI-based solutions should strengthen solutions and should be exploited as widely as possible. New frontiers of development should be considered.
The real challenge for everyone, not just IT specialists, is that the building blocks of KMS should form a real and viable system in the corporate practice. The AI solutions used to support the system should be integrated in each step to achieve a cooperation based on common principles, built on the same or similar logic, enabling them to be integrated into a workable system. Each of the steps presented cannot be explained without systematization. Most importantly, a solid framework for correlationships is essential.
References
Anklam, P., and S. Higgison. 2005. The Social Network Toolkit: Building Organizational Performance through Collaborative Communities. London: ARK Group.
Az öntanító AI. https://.groupmind.co/az-oentanito-ai
Badinszky, P. 2008. “BLOCKS—Személyes Tudásmenedzsment És Tartalomfejlesztő Szoftver.” Impresssum 10, no. 6. http://ujsag.szie.hu/node/542
Barkovics, B. 2016. “Tisztázzuk az Ipar 4.0 alapfogalmait!” https://autopro.hu/trend/Tisztazzuk-az-Ipar-4-0-alapfogalmait/18073/
Bencsik, A. 2013. Best Practice a Tudásmenedzsment Rendszer Kiépítésében. London: Pearson.
Biró, K. 2017. “A Robotok és a Mesterséges Intelligencia Lendíthetik Fel a Japán Gazdaságot.” https://sudy.co.hu/news/a-robotok-es-a-mesterseges-intelligencia-lendithetik-fel-a-japan-gazdasagot/
Birzniece, I. 2011. “Artificial Intelligence in Knowledge Management: Overview and Trends.” Scientific Journal of Riga 46, pp. 5–11.
Chena, Y.-H., T.-P. Lina and D.C. Yen. 2014. “How to Facilitate Inter-organizational Knowledge Sharing: The Impact of Trust.” Information & Management 51, no. 5, pp. 568–78.
Davenport, T.H. 1996. “The Future of Knowledge Management.” CIO 9, no. 5, pp. 30–31.
Davenport, T.H., and P. Laurence. 2001. “Tudásmenedzsment Budapest: Kossuth Kiadó.”
Duhon, B. 1998. “It’s All in our Heads.” Inform 12, no. 8, pp. 8–13.
Fehér, P. 2003. “Tudásmenedzsment: A Jövő Szolgáltatása.” In Szolgáltatások a harmadik évezredben, ed. P. Ilona, 417–62. Budapest: Aula.
Gholami, M.H., M.N. Asli, S. Nazari-Shirkouhi, and A. Noruzy. 2013. “Investigating the Influence of Knowledge Management Practices on Organizational Performance: An Empirical Study.” Acta Polytechnica Hungarica 10, no. 2, pp. 205–13.
Hanako, A. 2016. Artificial Intelligence and Knowledge Management. U.S.: Willford Press.
Hasznics, M., and J. Nuridsány. 2006. “Hálózati Tudásmenedzsment Rendszerek Sajátosságai, Bevezetési Tapasztalatai.” http://nws.niif.hu/ncd2006/docs/ahu/032.pdf
Hlács, F. 2016. “Együtt Kutatják a Mesterséges Intelligencia Hatásait a Tech-Óriások.” HWSW. https://.hwsw.hu/hirek/56225/partnership-on-ai-mesterseges-intelligencia-google-ibm-amazon-microsoft.html
Hoeschl, H.C., and V. Barcellos. 2006. “Artificial Intelligence in Theory and Practice.” In IFIP 19th World Computer Congress International Federation for Information Processing, Volume 217, ed. M. Bremer, 11–19. Boston: Springer.
Kiss, M. May 2017. “Tíz Eszköz, Ami Miatt a Mesterséges Intelligencia Nem Scifi, Hanem Inkább Szexi” Forbes. https://forbes.hu/legyel-jobb/tiz-eszkoz-ami-miatt-a-mesterseges-intelligencia-nem-scifi-hanem-inkabb-szexi/
Kulaklia, A., and S. Mahonyb. 2014. “Knowledge Creation and Sharing With Web 2.0 Tools for Teaching and Learning Roles in so-Called University 2.0.” In Procedia—Social and Behavioral Sciences 10th International Strategic Management Conference 150, pp. 648–57.
Lovászy, L. June 2017. “A Mesterséges Intelligencia Egyes Kérdéseiről.” Világgazdaság. https://.vg.hu/velemeny/mesterseges-intelligencia-egyes-kerdeseirol-532128/
Mi az OCR Technológia? 2017. https://ocrszoftver.hu/mi-az-ocr-technologia
“Mi is az a Mesterséges Intelligencia?—Mesterséges intelligencia II.” 2012. Infoter Marc. http://infoter.blog.hu/2012/03/07/mi_is_az_a_mesterseges_intelligencia_mesterseges_intelligencia_ii
Neururer, M. 2015. “Artificial Intelligence in Knowledge Management.” https://medium.com/artificial-intelligence-ai/the-role-of-artificial-intelligence-in-knowledge-management-309973209cfd
Nonaka, I., and H. Takeuchi. 1995. The Knowledge-Creating Company: How Japanese Companies Create the Dynamics of Innovation. New York, NY: Oxford University Press.
Ojala, M. 2016. “Best Practices in Cloud Solutions for the Enterprise.” KMWorld 7/8, pp. 14–21.
Poór, J. 2010. Menedzsment-Tanácsadási Kézikönyv. Budapest: Akadémiai Kiadó.
Probst, G., S. Raub, and K. Romhardt. 2006. Wissen Managen, Wie Unternehmen ihre Wertvollste Ressource Optimal Nutzen. Wiesbaden: Gabler GmbH.
Stone, P., R. Brooks, R. Calo, and G. Hager. “Artificial Intelligence and Life in 2030.” One Hundred Year Study on Artificial Intelligence: Report of the 2015–2016 Study Panel. Stanford, CA: Stanford University. http://ai100.stanford.edu/2016-report
Tölgyes, L. 2016. “A Mesterséges Intelligencia Kamaszkora”. IT Business, Marc. http://.itbusiness.hu/Fooldal/rss_3/A_mesterseges_intelligencia_kamaszkora.html