How can MNEs gain the Competitive Advantage by effectively implementing
Knowledge Management through Artificial Intelligence?
IBM Watson Case Study
Cristiano Nico (B70743862/S3899322) 2/12/2019
Word Count: 15,000
Master’s Thesis in International Business and Management Newcastle University / University of Groningen
ACKNOWLEDGEMENTS
ABSTRACT
TABLE OF CONTENTS Acknowledgements………. 2 Abstract………... 3 Table of contents………... 4 1. Introduction………... 6 2. Literature review……… 10 2.1 Knowledge Management………...……… 13
2.2 Knowledge Management and Globalization……….… 15
2.3 Knowledge Management and Artificial Intelligence……… 17
2.4 Ethical Aspect of Artificial Intelligence and Knowledge Management………... 20
2.5 Business Results and Future Perspectives of AI applied to KM………... 21
3. Methodology……….……… 3.1 Plan……… 24 25 3.2 Design……….. 26 3.3 Prepare……… 28 3.4 Collect………. 28 3.5 Analyze……….……… 32 3.6 Report……….. 34 4. Findings……….…… 36 4.1.1 Main Strategies………..………. 36
4.1.2 Features: Company Perspectives and People Perspectives…...……….……… 37
4.1.3 Outcomes: Business Perspectives and Knowledge Perspectives………….……….. 37
4.2 Exploratory Research Propositions………...……… 39
5. Discussion……… 52
6. Conclusion………..……. 57
Bibliography……….. 59
Appendix………..………. 75
Appendix A. Semi-structured Interview Questionnaires ……… 75
Appendix B. Semi-structured Interview Transcripts ……….… 78
1. INTRODUCTION
The IBM Watson case study presented in this dissertation aims to extend the literature on the application of AI in the context of KM. This manuscript aims to explore how a leading MNE leverages AI as a tool for everyday use. AI processes applied to KM are no longer thought of as a desirable future, but are a present, applied, and measurable reality. Through a set of semi-structured interviews addressed to IBM SMEs, the investigation seeks to analyze the processes and practices of AI employed by a large MNE that makes use of them worldwide. A 2019 report from the International Data Corporation (IDC), a U.S. provider of market intelligence research and advisory consulting, estimates that global growth in AI peaked at 35.8 percent as of 2018, with IBM holding the lead in market shares at 9.2 percent (Jyoti et al., 2019). The IDC report highlighted the spillover effects of IBM Watson, the company’s AI application solutions, into a variety of domains, spanning from agriculture and manufacturing to human resources (HR) management and marketing communications. As a result, the research highlights the main features that drive companies towards a digital transformation in the field of KM.
dissertation expands the current literature on AI applied to KM by investigating the experience of a leading MNE that has applied AI for enhancing intellectual capital and generating and disseminating new knowledge derived from the use of IBM Watson.
KM is crucial for business success. Smith and Farquhar (2000) assert that the main ambition of KM is to enhance organizational performance by empowering individuals to acquire, exchange, and implement their shared knowledge in order to achieve optimal real-time decisions. Carrillo et al. (2000) have extended this definition by suggesting that the purpose of KM is to identify, optimize, and manage intellectual resources actively to generate value, boost productivity, and obtain and maintain a competitive advantage. The dissertation will borrow Sigalas et al.’s (2013) definition of competitive advantage: “the above industry average manifested exploitation of market opportunities, neutralization of competitive threats and reduction of costs.” Knowledge is an intellectual resource: in the global economy, intellectual capital is overcoming traditional capital and labor assets as a crucial resource in developed economies (Edvinsson, 2000). The success of organizations in the current global and interconnected economy depends on how they can cope with rapid, effective, and efficient sharing of information (Kumar et al., 2014). However, as they grow and expand, it becomes a burden for them to deal with processing large volumes of organizational knowledge. As a result, businesses have to devote considerable attention, time, and effort to implementing effective KM processes. Combining AI to KM practices overcomes the amount of pressure arising from handling a large amount of information, thus providing the organization’s workforce with valuable information.
communications, and heterogeneous documentation. Every employee of a company is a potential contributor. KM instruments already incorporate some aspects of AI technology, such as intelligent agents, data mining, ontologies, as well as Bayesian reasoning. Besides, content management, personalization of human-computer interactions, user profiling, and case-based retrieval techniques are some of the many AI techniques available to be used in various aspects of core business processes as a result of Web-based technologies and components-based software engineering (Tsui et al., 2000).
Business scenarios are becoming increasingly global. Products and services offered worldwide by MNEs must face the challenges of multicultural differences. Information systems help people to make decisions, to solve problems, but more generally to integrate the contributions of people often located in scattered parts of the world. The development of IT solutions provides valuable support in accelerating the process of knowledge acquisition in a multicultural business environment. KM concerns organizational sharing of information and collaboration. Teams and groups face increasingly complex decisions. Managers that support group work and cross-cultural teams, where team members may work in multiple locations and at different time zones, need to take into account communication issues, technology-mediated cooperation, and work methodologies.
2. LITERATURE REVIEW
Up until the early nineties, the literature on AI applied to KM was very limited. Many companies lost interest over the potential of AI during this period. With "AI Winter," Crevier (1993) indicates a time-period when organizations felt that high expectations on the subject would lead to great disappointments. This has also meant a setback in research funding in this area. Carbone and Kersberg (1993) proposed the development of an automatic system that would facilitate the database interface in order to obtain a better data analysis. From 1993 to 2011, the field of AI and the related literature related to KM reborn, both because of technological evolution, through the use of cheaper but also much more powerful computers, and the desire to apply new scientific discoveries to different industrial fields. This phase can be defined as the "AI Dream," where computers can participate actively in economic development.
By the very end of the 20th century, AI began to receive critical acclaim from the public. In 1997, IBM supercomputer Deep Blue defeated Garry Kasparov, the world’s renowned chess champion. Millions of viewers followed the event around the world (McCorduck, 2004). In 2005, a Stanford University robot car won a U.S. competition by driving without human aid for seven hours along a desert trail in southwest Las Vegas. Machines were able to drive fast and safely without human intervention (Orenstein, 2005). In 2011, Watson, IBM’s AI engine, beat the two most celebrated champions of the American quiz show “Jeopardy!” by a significant margin (Markoff, 2011). The event marked a big step forward for AI, as machines were able to understand, react to, and potentially substitute human beings. Applied to KM, AI adds value to knowledge by analyzing and simulate human functions (Hoeschl and Barcellos, 2006).
Through AI implementation, organizations can leverage the breadth of their knowledge. Distributed Artificial Intelligence allows the acquisition of knowledge and semantic analysis of information obtained from the web (Gandon, 2002). Besides, AI tools and techniques not only allow for knowledge analysis and management but also to generate new knowledge (Liebowitz, 2001; Metaxiotis et al., 2003). AI systems identify the knowledge, acquire it, generate it, organize it, integrate it, and distribute it, thus improving the quality of organizations’ decision-making processes. The interrelation between AI and KM and the result of their useful application is the ability to learn and solve complex problems (Becerra-Fernandez et al. 2004). The effective use of AI in the business decision-making process concerns the results that intelligent systems allow achieving in terms of the position of advantage over an organization's competitors and better performance.
the positive effects of AI implementation in KM. Trends in the use of AI in KM focus on optimizing document management, research, and information sharing via blogs and wikis (Bizirniece, 2011). Mercier-Laurent (2014) explains how intellectual capital management is one of the main assets of today's organizations. The use of innovative techniques using AI can provide significant help in conserving, updating, visualizing, and searching for relevant elements of human capital. These systems will be one of the main levers for the transformation of companies towards digitalization and new forms of knowledge processing (Avdeenko et al., 2016).
Figure 2.1 Timeline of AI and KM literature 2.1. Knowledge Management
Numerous studies on KM (Drucker, 2001; Barclay and Murray, 1997; McAdam, 2000; Rosenthal-Sabroux and Grundesten, 2008; Dalkir, 2005; Nonaka and Van Krogh, 2009) allow us to understand how companies can transform primary data into structured information that can become useful work tools for their employees. The field of KM is established for more than 30 years and shifted from a vaguely defined concept to an essential element of organizational life. Over time, the nature of KM has evolved. Over the last decade, the challenge of determining an applied definition of the domain has moved from scholars to professionals. The 21st century introduced definitions of KM across a broad spectrum of disciplines (Girard and Girard, 2015). Jasimuddin et al. (2005) link KM to several disciplines, such as information systems, organizational analysis, strategy, and HR management. Various authors (O’Dell and Grayson, 1998; Davenport and Prusak, 1998) attempted to provide general definitions of KM, which emphasize the effective and efficient use of resources that enable organizations to improve their overall performance. Liebowitz (2012) claims that KM is the combination of three components: people (how to create KS environment and culture in the organization), process (how to manage KM processes and align the employees’ daily tasks), and technology (how to create a platform for communication and KS among employees).
well as exchanged (Tiwana, 2002). This form of knowledge is crucial for firms’ ability to collect, send, and even sell, and it can be stored in written and electronic form. On the other hand, tacit knowledge can be retained in people’s memories, and it is based on “intuition, feelings, faith, life experiences, and organizational culture” (Domagała, 2017). Woo (2004) defines tacit knowledge as the most critical asset for organizations seeking to gain a comparative advantage. In order to fully grasp its hidden value and to exploit the overall experience gained by individuals’ mental models over time, organizations must be able to convert tacit knowledge into explicit knowledge, a process known as externalization.
Tacit knowledge is difficult to codify, and externalization methods of a whole body of knowledge remain controversial. Johnson et al. (2002) argue that in the process of knowledge conversion from tacit to explicit, some of its original features may disappear. In terms of knowledge transfer and sharing within the organization, Polanyi (2009) identifies an array of processes that convert workers’ knowledge and tacit knowledge into valuable knowledge resources that allow an organization to gain a competitive advantage. Moreover, studies (Brown and Duguid, 1991; Hedlund, 1994) posit that tacit knowledge is context-dependent and is triggered and constrained by human relationships. McKinlay (2002) illustrates how a U.S.-based international pharmaceutical corporation developed an online archive to codify and disseminate tacit process knowledge beyond the single working group. Besides, research has linked the success of organizations in the field of technological innovation to their ability to leverage tacit knowledge gained over time (Seidler de Alwis and Hartman, 2008). Therefore, tacit knowledge is a critical resource that has the potential to sustain organizations’ competitive advantage and innovation.
controlling of people, processes, and systems in the organization to ensure that its knowledge-related assets are improved and effectively employed." KM is the entire workforce’s collective knowledge targeted at achieving precise organizational goals: it results in strategies and processes aimed at identifying, capturing, structuring, valuing, leveraging, and sharing the intellectual resources of an organization to improve its performance and competitiveness (Mohajan, 2017). KM systems can capture tacit forms of knowledge by externalizing and integrating them (Nonaka and Takeuchi, 1995). Davenport and Prusak (1998) focus on how organizations design processes that allow to capture, code, and transfer knowledge.
2.2 Knowledge Management and Globalization
At a later time, the KM literature was interested in how knowledge is created (Ocholla, 2011) and how organizations can apply it effectively to make decisions, innovate and create a competitive advantage over competitors in the marketplace (King, 2009; Bouthillier and Shearer, 2002; Gold at al., 2001; Wen, 2009; Birasnav, 2014). Significant theoretical developments (Marquardt, 1996; Jamali et al., 2006) addressed how companies can adapt their processes and technologies to transform information into applied knowledge. The world is continuously changing, and the two key elements that characterize it are globalization and the management of big data. In this case, the literature has addressed the problem through two separate studies: studies concerning cross-cultural implications (Maham, 2013; Ling, 2011) on KM; studies concerning the difficulty of collecting and managing big data (Marr, 2015; Hilbert, 2016) and the opportunity to obtain hidden information and weak signals.
researching diverse contexts, it is necessary to structure roles, responsibilities, and power between several organizational components, such as teams, units, and management structures (Del Giudice et al. 2012). Besides, identifying similarities and discrepancies in KS strategies of managers of diverse national and ethnic workforces is a crucial requirement for successfully designing flexible KM systems that can be adapted to the styles and needs of employees of MNEs around the world (Ardichvili et al., 2005). As a result, cultural contexts shape KM systems planning and implementation decisions, which must be in line with the different employees’ managerial styles and work values.
National cultures deeply intertwine with organizational behavior, which in turn affects KM decisions. As employees bring their individual culture to the organization through their different customs and language, organizational culture, in turn, impacts their values, behaviors, perceptions, and desires, including the willingness to sharing knowledge (Kreitner et al., 2008; Usoro and Kuofie, 2006). A study on information technology systems (IT) and KM supports the claim that business managers must adapt IT applications to the decision-making styles of people from different countries and cultures (Martinsons and Davison, 2007). What is more, research involving Chinese and American employees discovered that idioms, different mentalities, and different degrees of perceptual credibility of voluntary KS were three significant national culture differences impacting online KS in a multicultural context (Li, 2010). Therefore, KM professionals must analyze and understand the national and organizational cultural context in which they find themselves in order to apply knowledge correctly in the business environment.
2.3 Knowledge Management and Artificial Intelligence
Practitioners can now deliver more resources for training, quality control, and refining of AI results. Machines can increase the experience of their human counterparts and even assist in creating new specialists. These new systems, more closely imitating human intelligence, are becoming stronger than the large data-driven systems that preceded them. They could impact the 48% of the US labor force who are knowledge workers—and the well over 230 million roles of knowledge workers globally (Daugherty and Wilson, 2019). As a consequence, enterprises will ultimately have to redefine knowledge-work processes and careers to exploit AI’s potential. AI technologies have an essential role to play in the analysis and interpretation of all information obtained. This aspect requires particular consideration in light of recent experiences in the application of AI.
complex machine learning system. AI studies are often complex owing to the specific nature of the issues, which concern theoretical, practical, operational, philosophical, and ethical aspects. AI intertwines with autonomy and adaptability by learning from a dynamic environment (Miailhe and Hodes, 2017). Several studies (Strauß, 2018; Mangelsen and Alexander, 2019) deal with the application of AI to organizations that want to manage large amounts of data effectively. Tsui et al. (2000) explore how to make these massive amounts of data usable in terms of knowledge, cognitive, and predictive systems that facilitate decision making and problem-solving.
In order to apply KM successfully, AI technology must operate on a large population of people that Drucker (1999) defines as "knowledge workers" within companies. Daugherty and Wilson (2019) define knowledge workers as "people who reason, create, decide, and apply insight in non-routine cognitive processes." They defined knowledge-worker productivity as the most significant management challenge of the 21st century and the "first survival requirement" for developed countries. Without it, it would be unimaginable for them to maintain their leadership and living standards (Drucker, 1999). The organizations cannot learn or develop sound knowledge independently of their human capital (Bogdanowicz and Bailey, 2002). If knowledge workers act in line with the objectives of acquiring, sharing, and reusing knowledge, it is vital to understand how AI systems can address and solve problems through some autonomy of action and how it is possible to assist people in developing effective solutions.
an excellent solution for KM and BI: it is a computer system designed to address the needs of the future. Organizations must optimize heavy workloads in order to meet the new challenges of a future market that requires increasingly intelligent solutions. IBM Watson includes a set of industry-specific analytics solutions that leverage a new way to analyze cognitive content (Perrone, 2011). Watson develops through coherent reasoning of information to speed-up and make better decisions, minimize costs, and optimize results.
Managing knowledge is essential in many organizational contexts, such as health care, education, finance, transport, energy. IBM Watson is subject to continuous evolution by world-renowned experts who regularly draw new knowledge from the domain of competence and help people make informed decisions faster (Saravanakumar, 2019). Watson collects information from a wide variety of data types, including unstructured data, without additional integration, enabling the processing of extensive data archives. Through Watson, companies can transform the way they manage knowledge sharing by exploiting forms of natural language and generating hypotheses and new forms of learning. Watson combines several processing technologies and parallel probabilistic systems to improve the way companies solve problems. According to an IBM Watson document (2012), IBM's vision today is defining, establishing, and guiding markets towards innovative cognitive systems. These systems may be particularly useful where conventional approaches no longer work, the development of a cognitive class fosters secure and scalable modular solutions, and where the generated customer value is evident, demonstrable and quantifiable (IBM Corporation, 2012).
technology to help improve relationships, communications, and the use of knowledge by accelerating the dissemination of information and decision support.
2.4 Ethical Aspects of Artificial Intelligence and Knowledge Management
Ethical dilemmas vary across cultures, religions, and beliefs. Nevertheless, organizations can develop acceptable ethical frameworks to guide the reasoning and decision-making of AI technology to account for their actions. Governmental institutions and law-enforcement agencies contribute significantly to ensure that businesses and organizations adhere to and enforce their code of ethics. Last year, the European Commission appointed 52 representatives from academia, industry, and civil societies to form the “Independent High-Level Expert Group on Artificial Intelligence.” Recently, the AI expert group published fundamental ethical guidelines for ensuring “Trustworthy AI” (Independent High-Level Expert Group on Artificial Intelligence, 2019). Trustworthy AI systems should abide by all applicable laws and regulations and adhere to ethical standards and morals. Meanwhile, AI stakeholders and practitioners must acknowledge any unintentional harm that AI systems can cause. Therefore, the AI expert panel advanced four ethical principles for developers, deployers, and customers dealing with AI systems: “respect for human autonomy, prevention of harm, fairness, and explicability” (Independent High-Level Expert Group on Artificial Intelligence, 2019). AI technology must be adaptable enough to undergo regular updates and improvement as organizations identify and address ethical challenges.
experience. An organization that adopts AI systems in compliance with ethical principles must demonstrate transparency and reliability to the organization, its employees, and customers (Morgan, 2017). Undeniably, people and organizations can use implement AI technology to enable human self-actualization, fostering human agency, as well as enhancing social capabilities and cohesion (Floridi et al., 2018). Ethics applied to AI gives organizations a competitive advantage for recognizing and undertaking new and rewarding socially acceptable opportunities. Ethics also enables organizations to identify and prevent, or at least minimize, socially undesirable actions.
Successful organizations create and acquire new knowledge and use it to improve their operations and services. Organizations and personnel implementing ethics in their best practices can speed up quickly the conversion of explicit into implicit knowledge and vice versa. Both employers and employees face ethical dilemmas. Employers can misuse employees' knowledge without giving them credit for pooling the know-how. On the contrary, employees may withhold or divert the knowledge of their employer or team for their personal gains. Other ethical dilemmas concern the company's rights to limit access to knowledge and society’s rights to share organizational knowledge for the common good (Land et al., 2007). Rezaiian and Ghazinoory (2010) highlighted the relationship between integrity, mutual respect, trust, accountability, empathy, commitment, and KM processes. A more recent study reports that confidentiality, intellectual property, trust, confidence, and care in authenticity is of utmost importance in encouraging employees and organizations shifting from explicit personal knowledge to group and explicit organizational knowledge (Akhavan et al., 2013).
2.5 Business results and future perspectives of AI applied to KM
rethinking strategies, relying on a robust digital foundation, and develop integrated AI systems (Bughin and Hazan, 2017). A study by Capgemini illustrates similar trends. The report highlights the rise of “Smart Factories,” which are companies that use AI tools and can add up to $1.5 trillion to the world economy through digital transformation: the report registers an overall efficiency growth annually over the next five years, reaching seven times the growth rate since 1990 (Capgemini, 2017). The literature agrees that there is a continuous drive towards the company’s digitalization, a process that starts with a strong strategic vision and is implemented in a pervasive way in every aspect of the organization. Kruhse-Lehtonen (2019) argues that business leaders must create a business environment that supports digital transformation by paying attention to how they train their people, setting attainable organizational goals, and making substantial investments.
Managers willing to embrace AI and digital transformation must spread their message across the whole organization. In order to achieve significant business results, organizational decision-makers must communicate actively with their AI teams and stay abreast of technological improvements (Moldoveanu, 2019). In this way, it will be easier to define a robust strategy and a straightforward vision for the organization. A McKinsey survey projected that early adopters of cognitive technology benefit from higher economic growth than non-adopters (Chui et al., 2018), and Bughin (2018) warns that companies that are not investing in AI could lose competitiveness in the market. As a result, the literature highlights the importance of communicating decisions related to the digitalization and implementation of AI systems at all levels of the organization is essential for gaining a competitive advantage in the long-term.
3. METHODOLOGY
The research analysis follows a qualitative methodological approach. An inductive approach will be adopted to measure the effectiveness of KM systems that use AI techniques in MNEs. The study will proceed through the observation of specific cases applied to real scenarios and will obtain results that will allow the development of theoretical propositions. The research design will, therefore, be exploratory: the dissertation aims to explore the phenomenon under study, and the qualitative method adopted is the case study approach. Simmons (2017) argues that case study research is highly flexible and versatile. The case study methodology is an approach that guides experiential observation. The choice of the use of case studies to analyze issues related to AI and KM is particularly useful in deepening context-dependent knowledge and experience (Flyvbjerg, 2006). The case study approach allows examining the data within a specific context carefully. The aim is to consider one MNE, deepening the analysis by collecting experiences through the review of the documentation collected and through direct interaction with SMEs of the company involved.
The methodological application of the Case Study follows the indications of the social scientist and President of the COSMOS Corporation, Robert K. Yin, reported in the text "Case Study Research: Design and Methods" (2018). The organization of the case study develops through a linear but iterative path characterized by 6 logical steps:
1. PLAN: Understand if the Case Study is an appropriate research method;
2. DESIGN: Identify the cases to be analyzed and which type of case study will help to achieve the best results;
3. PREPARE: What to do before starting to collect Case Study data;
4. COLLECT: Collect the most appropriate sources that best fit the case study;
6. REPORT: Define how to represent the information collected concerning the purposes of the dissertation.
3.1 Plan
The use of the case study is highly complementary to regular scientific research (Eisenhardt, 1989). When deciding whether or not to use the case study approach, it is crucial to consider the type of question the research wants to answer, the level of control that the interviewer has over the behavioral events examined, and the events' development, focusing on the historical analysis or the recent past, and its evolutionary perspective (Yin, 2018). Besides, Yin (2018) argues that "Who, What, and Where" are issues sought through documentation, archives, investigations, whereas case studies require a more in-depth and detailed investigation that addresses “How and Why" questions. The research question for this case study analysis is the following: “How can MNEs gain the Competitive Advantage by effectively implementing KM through AI?”. The present exploratory research deals with a subject area that is little dealt with in literature, difficult to quantify, and characterized by thematic issues that require more in-depth investigation.
The case study analysis is then also appropriate because the research deals with contemporary events in which the researcher cannot manipulate specific behaviors during the investigation. The study will be exploratory and follows the evolution of the phenomenon over time instead of merely measuring its frequency, as in the case of historical statistical analysis. In carrying out the case study, data from different sources, such as documents and interviews, will be taken into account. The number of units examined by the case study is more limited compared to other research methods, such as surveys. Therefore, it is necessary to identify which experts in the field have gained adequate knowledge and experience to respond to and better describe specific events.
being related to observable and objective events. Observing how an MNE carries out its KM by leveraging AI is an objective element that the researcher can only detect without being able to influence it. Consequently, the orientation of the case study will be "realistic" because it describes a single reality independent of the observer (Yin, 2018). As recalled by Schramm (1971), a case study aims to enlighten a particular decision or multiple decisions.
When it comes to case studies, documentation, interviews, and secondary analysis are the primary sources of data. Researchers are encouraged to make greater use of documents, interview the right people, and make observations more unbiased (Yin, 2018). Moreover, addressing a particular audience and focusing on critical decisions will help to focus on the direction of the case study. The study examines the decision of some MNEs to use AI techniques in KM: why they make this decision, how they implemented it, and what corporate benefits they bring in terms of competitive advantage. Besides, the case study involves the triangulation of data from different sources of evidence.
3.2 Design
researcher would have spent a considerable amount of time contacting SMEs of other companies, thus jeopardizing the overall focus on important issues related to AI and KM.
To answer the research question and develop propositions, it was decided to investigate the case in-depth. The study examines a holistic, single case study. In order to evaluate the research project’s overall quality, Yin (2018) suggests the application of construct validity, internal validity, external validity, and reliability:
TESTS Case Study Tactic Phase of Research in
which Tactic Occurs Construct
Validity
Use multiple source of evidence Have key informants review draft case study report
Data collection Composition (case study
final report) External
Validity
Theory in single case study Replication logic in multiple case study (not used)
Research design Research design
Reliability Use Case Study Protocol Develop Case Study Database Maintain a Chain of Evidence
Data collection Data collection Data collection Table 3.1 Case Study Tactics (Yin, 2018)
all the information collected and complete with the researcher's report. As a result, the research guides the reader through the case study from the initial research question to the case study discussion.
3.3 Prepare
Yin (2018) advises researchers to abide by critical practices in order to prepare the case study adequately. Case study analysis requires them to know how to ask relevant questions and interpret the interviewees’ input correctly. The research sets up questions at various levels. The first group of questions will be more generic and address several SMEs, whereas the second set of questions derive from analyzing the results of the first set and will allow for further inquiry. Besides, Yin (2018) asks researchers to be good listeners, avoiding preconceptions or existing ideologies. In order to capture large volumes of data without bias, active “listening” skills are applied not only to semi-structured interviews but also documentary evidence.
Yin (2018) argues that case study analysis demands to be adaptive, considering all situations as opportunities and not as threats: during the analysis and deepening of the different themes, an adaptation of the contents may be necessary to maintain an impartial and unconditional position. The case study researcher must develop a mastery of the issues dealt with and have a firm grasp of the relevant theoretical issues in order to make analytic judgments when collecting data (Yin, 2018). A careful inquiry of the subject matter is carried out through documents, websites, conferences, and online public domain interviews and speeches. Moreover, the study adopts appropriate ethical behavior and consideration by avoiding any bias and being sensitive to contrary evidence by developing ethical behavior.
3.4 Collect
triangulation aims to enhance the case study’s construct validity by providing multiple measurements about the same phenomenon, and the diversification of data sources allows for a broader and more complete development of themes (Yin, 2018). The research will organize and document all the data collected through the construction of a case study database, which will help to identify possible relationships, highlighting repetitive elements, and increasing the transparency of the results. The analysis has employed word processing tools (i.e., Microsoft Excel and Microsoft Word) to arrange the data. In order to increase construct validity, the research maintains a chain of evidence. The findings have narrative materials derived from the case study database, referring to interviews and company documentation.
The researcher has selected the interview participants based on the skills, knowledge, and activities that they were able to experience in their work. The interview respondents are SMEs located in operating in the researched MNE’s offices located across Europe and use AI platform IBM Watson daily. For the most part, the interviewees worked in Italy (8), except for two SMEs working in the United Kingdom (1) and the Netherlands (1). The SMEs interviewed hold degrees in different fields of study: Computer Science, Industrial Engineering, Electronic Engineering, Biomedical Engineering, Physics, Economics, Organizational Theory, and Master of Business Administration (MBA). As for the documentation part, the research collected MNE’s reports, white papers, as well as public domain interviews and speeches available on the Internet with SMEs who hold, or have held in the past, prominent positions in the MNE and discussed the role of IBM Watson on work practices.
investigation. Also, the research will integrate data collected through semi-structured interviews with experts in the field. The research will examine one of the nine MNEs reported by Forbes that exploit AI technologies’ potential (McKendrick, 2019). Information will be collected and processed based on how these companies use AI and generate knowledge, how knowledge is shared and becomes a company asset, and what results businesses have achieved through the application of these tools.
The data will be processed to describe the different phenomena and lay the foundations for the next phase of the study. The choice of a large MNE derives from the peculiar characteristics and experiences gained by this company in recent years, both on the subject of KM and the use of advanced AI techniques. The data was collected using multiple methods, analyzing different sources, in order to triangulate the results. The data collection follows a theoretical sampling developed in two phases: in the first phase (first level), the research interviewed SMEs that had a broad view of the topic and collected documents to obtain data to cover the whole spectrum of the research question. In the second phase (second level), the analysis has deepened specific areas and researched data in order to confirm or modify the categories of the developed theory. The data collected were organized through "sensitizing concepts" (Bowen, 2006), that is, guiding principles that represent the starting point of the research. The research mainly asked open-ended questions (What? How? Why?) to encourage the development of divergent thinking by respondents.
Quantitative Details of Interview and Document Data
Source description Level 1 Level 2
IBM AI Cognitive Delivery Manager (direct interview) X IBM Client Executive AI SME (direct interview) X IBM Technical Solution Architect Cloud & AI Cognitive (direct Interview) X IBM Senior Managing Consultant & Research Scientist IBM Watson AI & Advanced Analytics (direct interview)
X
IBM AI Cognitive & Analytics Consultant (direct interview) X
IBM Senior Watson AI Consultant (direct interview) X
IBM Europe Automation Practice & Delivery Leader – AI SME (direct interview) X IBM AI IBM Watson Explorer Architect - IBM Analytics Europe (direct interview) X IBM Information Technology Architect – AI IBM Watson Dev Squad Team (direct
interview)
X
IBM Project Manager Application Automation (direct interview) X IBM Chairman, President and Chief Executive Officer (Ginni Rometty) (public video
interview)
IBM AI Ethics Global Leader, Distinguished Research Staff Member -- IBM Research AI (Francesca Rossi) (public video interview)
IBM Strategy & Operations Lead, MIT-IBM Watson AI Lab IBM Research (Mark Weber) (public video interview)
Former General Manager, IBM Watson Solutions (Saxena) (public video interview) Former Global Leader - Cognitive Visioning and Strategy - IBM Watson
(Bjorn Austraat) (public video interview)
Former Chief Technology Officer, IBM Watson Solutions (Sridhar Sudarsan) (public video interview)
Former Senior Vice President of IBM's Watson and Cognitive Solutions (David Kenny) (public video interview)
IBM Watson reports and white papers (7)
Table 3.2: Quantitative Details of Interview and Document Data 3.5 Analyze
knowledge in an impartial and unbiased manner (Yin, 2018). A logical model briefly describes the entities and relationships between the entities under investigation. The research conducted an inductive analysis of the data through techniques borrowed from grounded theory (Glaser and Strauss, 1967) and research into themes and forms of aggregation (Gioia et al. 2012).
The research has then proceeded to the “In Vivo” (Strauss and Corbin, 1990) coding of the identified and aggregated concepts. First-order coding (Van Maanen, 1979) uses single descriptive quotations to focus on critical concepts. As for data coding, the analysis explored the relationships between concepts in order to highlight the emerging framework. The technique used does not follow linear processes but recursive structures where the collected data are gradually refined and contextualized (Locke, 1996). The data structure, as shown in Figure 3.1, identifies the first-order key concepts, second-order key themes, and the three dimensions of aggregation (Strategy, Features, and Outcomes):
Figure 3.1: Data Structure
(see Appendix B), and the organization of data through key concept identification. The second action concerned the splitting of the data collected in two stages in order to ensure a first general view (level 1) and a deep dive on the topics under investigation (level 2). A third action concerned the comparison of the concepts found in the interviews with public domain documentation to detect any confirmations or discrepancies.
3.6 Report
In order to report case study results, the researcher will select the information to be included in order to highlight the most significant results. Also, a practical analysis of the results allows defining optimal forms of interpretation (Yin, 2018). The report will be developed throughout the course of the case study and will be organized based on the characteristics of the audience who may not be an IT expert. Therefore, the study will not go into technical details. The reporting will highlight how the study will contribute to enriching existing knowledge and developing new knowledge.
4. FINDINGS
This chapter presents the research results through the processing of the data related to the subject of AI applied to KM with particular focus on the experience gained by a leading company in the field of International Business and IT. The research highlights the results through a framework of synthesis and comparison with the main studies on the subject. Data analysis has enabled the researcher to relate the three aggregate dimensions and to identify a "Data Process" model (see Figure 4.1), which highlights the primary "Strategies" that drive companies and individuals to apply new "Features." The "Features" are those tools and techniques of AI that allow obtaining significant outcomes for the evolution of Business and KM.
Figure 4.1: Data Process 4.1.1 Main Strategies
data. These two main strategies, peculiar characteristics of IBM, are also common to many large companies that see these two elements as the key to their corporate vision. In order to implement these two AI strategies distributed on the Cloud, it is necessary to define how to accomplish these strategies.
4.1.2 Features: Company Perspectives and People Perspectives
The “Features” that define how to implement the two main strategies follow two perspectives. The “Company Perspective” feature represents the key elements that drive companies to use AI tools. The first element concerns the transition that companies make towards the digitalization of information and knowledge. The second element refers to the gradual integration of AI within all business processes. Lastly, the third element involves opening up towards innovation and reassessing operational methods and tools in every business process. The “People Perspective” feature represents the elements that impact on people’s activities, the tools they use, the behaviors they take, and the results they obtain. AI is no longer a mysterious and complex black box, but easy to use tools that improve the overall standard of living, and consequently, the performance of those who work with them. AI adoption improves decision-making processes and complex problem resolution by offering systems capable of gathering new needs and insights that would have remained hidden and unexplored without it.
4.1.3 Outcomes: Business Perspective and Knowledge Perspective
explained that through the use of Watson technologies, “the manager can make more use of knowledge and create more content." Besides, the interviews have touched upon issues related to cost savings. AI systems can perform faster and more accurate automated operations than human beings. The interviewees have assessed these aspects positively, highlighting a more efficient allocation of human resources. In other words, people will no longer engage in repetitive tasks but will be involved in more qualifying activities.
The "Knowledge Perspective" outcomes represent how AI revolutionizes KM through an innovative, effective, and continuously evolving approach. AI available on Cloud favors knowledge dissemination, as evidenced by several interviewees. An IBM Senior Managing Consultant claimed that "the computing capabilities of the hardware […] and the possibility of sharing them on the network through the Internet, […] the cloud itself, and […] the richness of statistical models and artificial intelligence that IBM develops for each case of application, are combined." An IBM report confirms that “organizations can handle structured and unstructured data in one platform, and they can capture and share models, dashboards, and notebooks. Data scientists save a significant amount of time on finding and preparing data" (Forrester, 2019). Former IBM Watson Solutions’ General Manager, Manoj Saxena, asserted that AI could manage the “knowledge curve for humanity,” allowing machines to capture and store experts’ current knowledge and experience so that future generations can learn and benefit from their insights (TEDx Talks, 2013).
"Personal data can be controlled completely. AI with IBM Watson looks at who needs information, then if the person has it in excess, then what is the level of content that the person has, and what is the time frame for which that information needs to be provided.” (Interview with IBM Europe Automation Practice & Delivery Leader, 2019). The framework below outlines the most critical quotations collected from direct, semi-structured interviews, as well as external documentation, grouping them into key concepts that summarize and interpret the data collected
4.2 Exploratory research propositions
Figure 4.2: Step 1 – Defining a robust strategy
IBM has defined a clear strategy of developing AI on distributed cloud networks. IBM’s vision is to develop a centralized platform of integrated services accessible by anyone on the Cloud through the AI platform IBM Watson. IBM Technical Solution Architect defined the nature of IBM as “a company founded on two main principles of information technology, one is the principle of the cloud, and the other is the principle of artificial intelligence." IBM has defined a strong corporate "vision" in which AI and Cloud computing are critical elements for business development. The strategies outlined by IBM represent the way to achieve the objectives set out in the vision.
having believed in advance in the transformation that was able to bring cognitive artificial intelligence to the time of machine learning, gave us a very good advantage in competitive terms.” A strong belief in its AI and Cloud strategy allowed IBM to position itself “a couple of years ahead” of its competitors, as deduced by IBM Cloud Executive.
In addition to AI, IBM's strategies are moving towards the use of cloud computing, which allows data dissemination and advanced cognitive analysis tools to anyone, anywhere in the world. Cloud computing is a revolution in the field of an industry historically very tied to the strength of its brand. IBM’s cloud and hybrid cloud strategies are to develop “a platform that is both distributed to [IBM’s] customers and to [IBM’s] centers,” in which everyone’s knowledge is critical “to create the best possible service for our customers” (IBM Cloud Executive). The acquisition of American software company Red Hat, Inc. by IBM proves the abovementioned statement and the words of its CEO, Ginni Rometty, confirm it. Rometty contended: “We actually had to do a lot of work around the IBM cloud private which is what Watson runs on […] Red Hat is coming up and so this allows it to move anywhere out there. This is a big piece […] of hybrid cloud which you've heard me say we think that's a trillion-dollar market and we'll be number one in it so that gives you a good feeling” (CNBC Television, 2019). A strong strategy requires substantial investments, even in times of economic crisis.
sacrificing accuracy” as IBM Strategy & Operations Lead, Mark Weber claims in a public domain interview (RE-WORK, 2018). In addition, John E. Kelly III, Senior Vice President and Director of IBM Research argued: “AI systems […] will require new innovations to tackle increasingly difficult real-world problems to improve our work and lives” (IBM Corporation, 2017). Through this ambitious strategic partnership with MIT, IBM aims to explore the economic and social benefits of AI in advancing knowledge acquisition to tackle societal problems and improve the human condition.
PROPOSITION 2: AI implementation orients companies towards digital
transformation, changing business processes, accelerating knowledge dissemination and sharing, thus benefiting from better use of intellectual capital.
Figure 4.3: Step 2 - Drive Company Transformation
IBM Watson integrates into workflows by adding AI where it is needed.” In order to upgrade and improve business processes, organizations must change their internal processes, by identifying focus areas where to implement AI technologies and take action in terms of efficiency, speed, and accuracy. Organizations need to understand which workflows, resources and, above all, the skills they need to be able to use AI tools effectively and trigger business optimization.
Digitalizing and automating workflows enable to redesign business processes through new information management, thus providing a different way of conceiving day-to-day operations. IBM Client Executive describes the IBM Cloud environment as “an infrastructure enabling the company to make a transformation to the digital world.” At the same time, digitalization is a trend that is affecting gradually many organizations. An IBM White Paper (2019) highlights that over the years, "firms are embracing more data sources on the cloud, combining it with existing data on-premises, and applying analytics and AI on the Cloud to drive new insights.” Companies oriented towards digital transformation understand that data digitalization and structuring add value to their core business because people can take directly the information needed, as IBM AI Cognitive Delivery Manager explains. Digital transformation also helps organizations speed up document management processes by focusing only on the most important data (IBM AI Cognitive & Analytics Consultant) and “perform[ing] analytics on […] large datasets to understand which dataset is corresponding to another, adding more insights” (IBM Europe Automation Practice & Delivery Leader). Cloud and hybrid cloud strategies significantly facilitate access and analyze large
Cloud & AI Cognitive described the role of the data scientist in business process optimization by minimizing costs and ripping the benefits of efficiency by exploiting the platform’s algorithm. Therefore, data scientists “can find those innovations and those steps that allow us to make certain business processes more effective and more efficient”, and by analyzing various users’ behavior, they can “understand how to improve and predict further action.” As data scientists can now access, use, and analyze larger data sets, their contribution to companies’ strategic processes improves remarkably.
PROPOSITION 3: AI applied to KM helps people make better and more objective decisions, allows them to solve complex problems better, and improves their work performance.
Figure 4.4: Step 3 – Reinforce People Transformation
the system “learns” and “improves” over time, overcoming challenging business problems, and transforming them into opportunities. An IBM Senior Managing Consultant and Research Scientists argues that when collecting a large amount of data, "in encoding its relevance to the specific decision-making domain and in allowing also a human understanding […] a greater decision-making capacity is allowed, because […] [data is processed] according to classifications that are then screened by the experience of managers and [SMEs]."
IBM Watson allows people to make more sound and effective decisions to complex everyday challenges, which leads to improved work performance. An IBM Senior Watson AI Consultant argues that "the decisions of the professional [are] more facilitated by more information." KM processes improve, because the AI system allows people to “access to unstructured data, and can learn from small data sets, […] and helps to increase its value by analyzing it more deeply,” says an IBM Project Manager. The research exposed these advantages not only at the corporate level but also in other domains, such as the medical field. David Cole, IBM Watson Health Innovation Lead for Europe, in a conference at the Oxford Union, discussed the significant contribution of AI to medicine (OxfordUnion, 2016). Rob High, IBM’s Vice President, argued that “by showing where the information and recommendations are coming from, Watson expands what human doctors can do and provides them with resources to make the best decisions for their patients” (Morgan, 2017). AI thus enriches the wealth of knowledge of doctors, helping them to make more accurate decisions in a shorter time frame.
only basic IT knowledge and brief training. An IBM Watson Explorer Architect argued that “it [is] a question of getting comfortable with them and the greater challenge is giving accurate and effective data to train them." Some SMEs argued that the AI platform’s simplicity of use depends on the type of application. An IBM Senior Watson AI Consultant argued that “some [applications] can be used even if you do not have specific knowledge, and then it is enough, others instead require the technical knowledge.” Nevertheless, the simplicity with which people can access and use of AI systems in cloud networks eliminates any psychological barriers and drives people to accept change, experience it positively, and change their habits.
PROPOSITION 4: AI-augmented business processes and people's behaviors transform and improve KM in terms of information collection and dissemination, data processing, and insight generation while ensuring data privacy and protection.
Figure 4.5: Step 4 – KM improvements
that AI techniques help people to seek more-in-depth knowledge and insights, thus helping people to extract concepts “at 360 degrees.” An IBM Senior Managing Consultant and Research Scientist deepened this concept by clarifying that the AI system recognizes “insights that the same human being accomplishes, but that then struggles to put together in correlation between […] thousands and thousands of records of data." KM has changed dramatically over the years. People interact with AI systems, organizing the information, and disseminating it quickly and pervasively. MNEs that have understood the importance of these processes have equipped themselves with cutting-edge tools to respond to new challenges in order to achieve new and more ambitious goals.
IBM is implementing AI tools and capabilities systematically in the evolution of KM processes, unleashing new ideas, and opening up new business opportunities. However, the quality of the processed data may not always be adequate, and the training processes of the learning machines may not provide clear, correct, or updated guidelines. Besides, the processing of emotional signals, feelings, and perceptions may not lead to objective evidence but may be conditioned by contingent factors. This is one of the issues addressed with the SMEs surveyed during the second-level interviews. When asked about any adverse effects deriving from incorrect or obsolete information that may arise when teaching IBM Watson, several interviewees confirmed the presence of these potential risks. For instance, an IBM Watson Explorer Architect argued that an effective governance system must be in place and that end-users must decide which approach (e.g., machine learning, linguistic rules) is the most appropriate for the training depending on the situation. On the contrary, IBM Europe Automation Practice & Delivery Leader did not highlight any possible dangers when training AI systems: “Watson is constantly learning. So even if the information is incorrectly input and coded into Watson, it will be quickly rejected.”
Delivery Leader ensured that "personal data can be controlled completely.” He also added that “Watson looks at who needs information, then if the person has it in excess, then what is the level of content that the person has, and what is the time frame for which that information needs to be provided […] all of those things can be deployed to effectively make sure of compliance to all regulatory bodies." Besides, an IBM Watson Explorer Architect added that “if [a person does not] wish, IBM will not learn from that data." The direct interviewees, as well as public domain documentation, expressed optimism and trust towards IBM’s data protection and privacy, and the company’s full transparency towards data processing. Francesca Rossi, IBM AI Ethics Global Leader, argued in a public interview that an AI system is trustworthy when it “is not biased, is fair, is explainable, and the way uses the data of the user is transparent” (ITU, 2018). Also, she guaranteed that IBM does “not reuse the data for other clients or other tasks” (ITU, 2018).
navigating through concepts, making the search more aggregated and finer-grained, so that people can attain the intrinsic value of that particular type of information.
PROPOSITION 5: Business transformation and new ways of managing knowledge
through AI capabilities lead companies to achieve positive results in terms of generation of innovative insights, resulting in increased revenue, cost reduction, and resource optimization.
Figure 4.6: Step 5 – Improved business performance
domain talk, Manoj Saxena discussed the benefits of the AI platform in terms of empowering the way people think, act, and learn (TEDx Talks, 2013).
Some SMEs and publicly available documentation have highlighted the benefits of AI in HR processes, which provides faster and more accurate information analysis, thus speeding up personnel management systems. IBM Technical Solution Architects posited that AI “helps to find the right match between people's skills and the job”, and contended that automatic systems help people to “trace [employees’] professional evolution”, suggesting them “what are the most appropriate things […] to put in [their] curriculum […] on the basis of real evidence." In a public domain interview for the University of California, Berkeley, former IBM Global Cognitive Visioning and Strategy Leader Bjorn Austraat draws a link between cognitive transformation and HR transformation, arguing that AI allows for “a transformation of functions, individual functions, but then also of the overall enterprise […] from the complete employee and engagement lifecycle” (Berkeley Haas, 2017). An IBM document illustrates employee cost savings of AUD 10 million by an Australian energy company. These savings stemmed from the use of IBM Watson, which allows "faster access and more intuitive analysis of [...] records" leading to "a 75% reduction in team time spent reading and searching data sources" (Banerjee, 2017). By looking at the organization holistically, AI allows to improve business processes and models and reduce costs, allowing organizations to gain a competitive advantage.
5. DISCUSSION
The case study has explored critical aspects related to how MNEs can gain a competitive advantage in the implementation of KM practices through AI. Through the deepening of a real and practical experience of an MNE that regularly uses AI techniques and capabilities to manage its intellectual capital, the research highlighted the positive impacts of these best practices on people and business results. AI is transforming the way information is collected, processed, and distributed. This has led to the generation of new levels of knowledge that foster new ideas and new business opportunities. The discussion section relates the results obtained with the current literature on AI applied to KM.
The second proposition highlights how AI application in KM processes drives companies towards digitalization, improves the information dissemination, and yields clear benefits to organizations' intellectual capital. Existing research agrees that companies are making a transition to digitalizing KM processes by applying AI capabilities when processing large amounts of data (Paschek et al., 2017). As confirmed during the interviews and public domain documentation, the literature also illustrates how digital transformation drives companies towards digitalization and new forms of knowledge processing (Avdeenko et al., 2016). A report by Capgemini (2017) on the digital revolution illustrates how this aspect radically changes every business process.
The literature agrees that in order to obtain the best AI-driven business results based on AI, the action of data scientists is not enough. Digital transformation must take place throughout the whole organizational environment. Kruhse-Lehtonen (2019) discusses how digitalization can create greater efficiency and productivity: business leaders must set ambitious but realistic goals, look after people and give them adequate training, identify the most appropriate digital investments, and implement operational models to organize data and AI effectively. Managers, and companies more in general, benefit from their digital transformation not only through increased revenues and savings, made possible through production improvements but also through more effective use of data and knowledge (Botha, 2019). The IBM Watson case study illustrates how an MNE has responded to the need, already expressed in the literature, to integrate AI tools and techniques in its organizational processes. The exploratory research has confirmed that digitalization allows for greater dissemination of knowledge and the optimal use of intellectual capital.
analysis thanks to AI. Liebowitz (2001) discusses the importance of AI in knowledge discovery and data mining approaches, which could be implemented inductively to find relationships in repositories for new knowledge creation. Indeed, many large companies have made investments in AI R&D a priority for their core business. Mortensen (2019) stressed the importance of developing emotional thinking and higher-order expertise to cope with the most sought-after skills of the future. Botha (2019) highlights how knowledge workers will need to re-skill their jobs, personalizing, and sharing contextualized knowledge in support of digital transformation. As a result, a more careful collection, selection, and enrichment of knowledge will allow them to make better decisions.
The fourth proposition describes how the implementation of AI can transform the way of managing business communication by processing vast amounts of information and generating new knowledge while preserving users’ data privacy. Gandon (2002) has focused on how AI can exploit the breadth of human knowledge, starting with information shared via the web. The world changes, and the amount of data being shared and analyzed increases. The application of AI in business processes enables large amounts of data to be managed effectively (Strauß, 2018; Mangelsen and Alexander, 2019). The change brought about by the use of AI in KM is addressed in the literature mainly in terms of opportunities: document management optimization, information sharing and research (Bizirniece, 2011), and support for people to make more informed decisions more quickly (Saravanakumar, 2019). Semi-structured interviews from IBM AI experts, as well as public domain videos and documentation, illustrate how the practical application of AI in KM helps people make better decisions and solve complex problems.
Although they claim IBM has equipped itself with all the necessary tools to ensure its stakeholders data protection and privacy, the literature on these issues is inclined to highlight scenarios that are not always positive and highlights the need to find tools and techniques that promote transparency and accountability in data-based decision-making (“Big Data Senior Steering Group”, 2016). Some articles consider the protection of personal data as a weak point that calls into question the security aspect of AI systems, as these may not apply the necessary control measures to protect customers and employees (Cate and Dockery, 2019; Shaw, 2019). Although they recognize the significant benefits of AI at the systemic, corporate, and individual levels, the authors point to severe gaps limiting data protection frameworks, which are inadequate to protect people's privacy and promote innovation in the data-based economy.
6. CONCLUSION
The study has deepened the experience gained by an MNE in the field of AI applied to KM. The research has shown that AI systems can improve considerably the way people collect, analyze, and share information. By applying innovative forms of human-machine interaction, companies can achieve positive results for their core business and their stakeholders’ quality of life. The case study allowed us to understand how KM has changed in an MNE that implements AI tools in its business processes: people fully exploit their intellectual capital, and AI systems generate new knowledge and enrich corporations' knowledge base. Computers do not replace people but integrate them and develop their potential so much so that today, companies refer to AI as "Augmented Intelligence" (Jablokov, 2019) that renovates decision-making processes and facilitates complex problems resolution. Through the IBM Watson case study, this dissertation offers a virtuous path that MNEs can follow. The model starts from the setting up of a robust strategy, a vision that can trace the desired future in which increasingly sophisticated cognitive systems improve the capture and management of knowledge within MNEs. The study also stressed that aspects of personal data protection cannot be underestimated and must follow particularly strict rules and forms of control.
systems, consider them an integral part of their work, and are not surprised to talk with an automatic chatbot to get the information they need.
This dissertation did not consider specific technical aspects of IT solutions that would have allowed the researchers to go into the operation of AI tools, advanced methods of machine learning, and possible future technological developments. Besides, this thesis did not consider quantitative aspects related to research, such as the amount of investment in R&D, the increase in revenue from new business opportunities, savings on personnel costs, return on investment. An area of possible future work relates to the measurement of quantitative aspects of positive business results stemming from AI implementation in KM, such as by calculating the ROI of AI (Return of Investment of Artificial Intelligence), as proposed by an Accenture research report (Mannar, 2019). Another point not covered by this manuscript concerns the analytical comparison with the experiences of other MNEs that have undergone similar transformations in the field of AI and KM. The comparison between MNEs could highlight significant differences, analogies, and consequent results on companies’ corporate strategy and human capital.
