**8. Conclusions and discussion**

Organizational management research has typically focused on qualitative behavioral factors that have a complex relationship to organizational success, and in addition, impacts often come with a delay. Each organization is a unique system with certain same laws, but also a unique context of its own. Therefore, repeating the empirical research results has proven to be challenging, which also makes it difficult to draw generalizable conclusions [7]. This article examines the utilization of model-based artificial intelligence in management development. ODT can be used to assess the impact of management behavior on an organization's success, considering situational data and the impact of management culture. ODT helps to explore

the fundamental nature of an organization, which means a metaphysical essence in where everything affects everything.

The article uses artificial intelligence to illustrate how leadership behavior can create a so-called QWL glass roof that invisibly prevents teams from growing to the top performing category. The management system forms the behavior of supervisors in such a way that harmful biases of management thinking may occur, in which case people's performance does not develop favorably. These harmful biases of thought are very complex as they include phenomenon of delayed effects on an organization's competitiveness. Model-driven reinforcement learning artificial intelligence reveals a variety of human and complex mechanisms that hinder the development of competitiveness.

Reinforcement learning is following rational learning phenomenon, where learning take place gradually, according the experience. Simulation model provides learning platform where person can learn without fear of remorse. This is essential especially for managers, because in real life there is hardly room for learning from mistakes. The ODT models the situation with the organization's own data. The simulation can be designed according to the company's own strategy, allowing future challenges to be practiced. This allows management and supervisors to adapt in advance and prepare for future challenges. More proactive management reduces the realization of personnel and business risks and adds value to performance. For example, adapting to a recession can be practiced, as can market growth, both of which require a different way of managing. Artificial intelligence combined with the digital twin helps to emphasize leadership skills and practices that lead to sustainable development.

ODT has been used in college students' leadership studies. Learning outcomes have been monitored through self-assessments, and the results are encouraging. Gamified simulation-learning is based on reinforcement learning, where progress takes place through experiential adaptation according to the student's capabilities and learning ability [25]. ODT is also used in managerial trainings for companies and municipal organizations. Perhaps the biggest challenge in coaching supervisors in working life is unlearning the biases that prevent leadership success. Traditional teaching is largely based on sharing best knowledge, where the teacher shares information on how to act and why to behave in a certain way. The power of digital simulation teaching is based on the fact that it adapts the brain through experiential learning. When a supervisor has to change the prevailing leadership attitude, he or she kind of adapts the brain to another frequency where listening and caring for employees rises higher in priorities. In this way, the supervisor becomes interested in developing herself in interaction practices where she may not have previously felt the need to learn.

The architecture of the digital twin models the reality of an organization with relatively good accuracy, which is important in building trust in an artificial intelligence solution. The core of the model is in the Human Capital Production Function of and in the scientific research of the Quality of Working Life index [26]. The architecture lays the foundation for a neural network that has been finetuned with the probabilities of empirical research as well as correlations created through supervised learning. For example, the physical and emotional safety (PE) of the QWL index correlates with sickness absence, so that when the PE factor falls, sick leaves increases. The correlation is brought into the digital twin, which makes the model more accurate because it also models sick leaves. In addition to research data, the digital twin can be calibrated with data from the organization. ODT learning can be extended in the organizational hierarchy to the level of an individual supervisor. In this way, artificial intelligence learns the strengths and weaknesses of a leader, so that the advice given by artificial intelligence is targeted at each supervisor.

**53**

**Author details**

University of Lapland, Rovaniemi, Finland

provided the original work is properly cited.

\*Address all correspondence to: marko.kesti@ulapland.fi

© 2021 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium,

Marko Kesti

*The Digital Twin of an Organization by Utilizing Reinforcing Deep Learning*

Supervised learning AI that is based on data alone is unable to "understand" organizational complexity and phenomenon of delayed impact relationships. In fact, there is a word of warning in using simple data-driven AI in complex organization environment, because it may strengthen the harmful behavioral biases. Article indicates that ODT with Bellman algorithm can be used in finding organization specific optimal behavioral patterns and measures which will form sustainable competitiveness. The article suggests that in the future, top-tier companies will use RL artificial intelligence to support management

*DOI: http://dx.doi.org/10.5772/intechopen.96168*

decision-making.
