INTRODUCTION
Individuals strive for democracy within every entity to actively participate in decisions that impact them, contribute meaningfully to the commu- nity they belong to, and assert their presence and importance (Bachrach, 1967, Butzlaff, 2023, Nabatchi, 2012). Although the forms of democracy may undergo changes over time, the inherent feelings elicited by democ- racy remain unchanged; even with the expansion of digitalization and its impact on democracy, questions of participation, representation, or trans- parency are becoming more decisive than ever (Stratu-Strelet et al., 2023, Pantazi et al., 2022). Digitalization also transforms how workplaces function, so while hierarchical organizations are typically not the ideal environment for democratic decision-making, stakeholders and managers of companies
© 2024. Published by AHFE Open Access. All rights reserved. 114
Democratization in Industry via Multi-Agent Systems? The Case of a Production Company 115
endeavour to pioneer the adoption of new technologies, fulfilling demo- cratic decision-making in workplaces (Charles et al., 2022, Hilton et al., 2021, Dingwerth et al., 2020). In the meanwhile, studies also acknowledged that democratic and participative approaches can effectively function within workplaces (Reuten, 2021); so, in this context, democratization via digital tools and its utilization in the workplaces opens up new possibilities and raises significant questions, demanding additional examination. Thus, this paper explores the question of democratization in companies using Multi- Agent Systems by analysing a case study conducted within a globally diverse manufacturing SME.
Driven by the aim of Democratization of Decision-Making in Socio- Technical Settings, our empirical investigation involved conducting inter- views with a selected cohort of eight employees from various divisions and hierarchical levels within the SME, facilitating a comprehensive exploration of multifaceted perspectives.
In this paper, dimensions of the decision-making process are first out- lined based on a case study, followed by the presentation of a concept for democratization via Multi-Agent Systems (MAS), and ultimately, some gen- eral remarks based on the findings as well as an outlook on our further research.
Methods
In this study, we employed a case study approach to examine the dynamics of democratization in industry through Multi-Agent Systems and to explore the contextual situation for implementing a Decision-Support System (DSS) within an SME (Yin, 2014). The case study focused on a comprehensive analysis of a specific production company, which initially involved veri- fying operational documentation describing workflow procedures through exploration of specified scenarios. The first scenario, ‘Automated test build- ing,’ begins with a new product order or the need for a production pro- cess update, facilitated by the program manager and forwarded to the automation engineer during the design phase. Another significant scenario, ‘Worker allocation,’ occurs prior to or during shifts, based on the assess- ment of workers’ abilities and the production plan. Additionally, in the ‘Machine maintenance’ scenario, triggered by breakdowns or failures dur- ing the execution phase, known problems are addressed through a ticketing system.
Our research encompassed first a document analysis, second a visit of var- ious departments within the company, beginning with an examination of company products and production halls, where ongoing tasks such as qual- ity checks, labelling, identification, and evaluation were looked at with the method of participatory observation. Subsequently, attention shifted to the operation of collaborative robots with human involvement. Third, we con- ducted interviews with employees across different hierarchical positions and levels of experience (N = 8). However, this represents only the first round of interviews, with plans for a second wave later this year. Moreover, workshop formats will be used to deepen the insights for the alignment of the results
116 Gheibi and Böschen
of the sociological analysis towards the development of the Multi-Agent Sys- tems. It should be acknowledged that the results reported here are based on the first step of our case study.
Understanding Decision-Making Processes for Developing Decision Support Systems To optimize the decision-support system towards a form of democratic decision-making, it is essential to closely examine the daily work routines of employees within the SME, along with their individual decision-making processes. To achieve this, eight interviews were conducted with staff across the organizational hierarchy, encompassing managerial personnel, supervi- sors and technicians from logistics and production teams, in addition to representatives from the workers’ council, inquiring about their decision- making situations, associated problems, and characteristics. Additionally, participants, whose experiences vary from 2 years to more than 10 years, were invited to discuss their involvement within the company, sharing their opportunities for objection and suggestion, as well as addressing the status of decision-making transparency and the significance of trust. Moreover, their expectations from a decision support system and potential functionalities to optimize its performance were explored. Thus, in the following sections, we investigate deeper into the insights gained from the interviews, focusing on three key dimensions: decision-making, involvement, and expectations, sub- sequently examining their respective categories. Each aspect offers valuable perspectives shared by participants, revealing finer details of developing a democratic decision support system within the organization.
Decision-Making Given that research indicates new technologies can enhance decision-making, it is essential for this study to obtain a clear understanding of the cur- rent decision-making processes within the SME (Rajagopal et al., 2022). This includes identifying whether decision-making follows a traditional or flat structure, evaluating the types of decisions made, assessing the effec- tiveness of the current process, understanding the challenges faced during decision-making, and recognizing the factors employees should consider in their decision-making approach. According to participants, decision-making situations can vary, ranging from decisions regarding the production line, such as assessing the authenticity of machine failure, determining when to report recurring errors to the supervisor, and potentially shutting down the line, to addressing stock differences or product shortages affecting produc- tion. These situations may also involve attributing fault, whether to internal parties, customers, or transport companies, in cases such as product failures and transport damage, and deciding whether to dismantle material returns. Moreover, decisions may involve personnel matters like overtime manage- ment, task allocation, and determining roles and timing for redundancy. Additionally, managerial and works council decisions may include adjust- ing budget allocations, managing staffing levels, handling dismissals, and addressing employee issues.
Democratization in Industry via Multi-Agent Systems? The Case of a Production Company 117
Based on statements from interviews, in the decision-making process, employees may face various challenges, with one of the most crucial being feelings of uncertainty regarding decision accuracy. This uncertainty may arise due to a lack of experience or insufficient information about the sys- tem, components, customer, employees, and timestamps. Consequently, the uncertainty presents itself differently: experienced individuals tend to accept the consequences, while amateurs often experience significant stress. Uncer- tain decisions can also jeopardize production line stability, particularly when made amidst disparities. These situations can become even more challeng- ing, especially when superiors or experts are unavailable for consultation or conversely when multiple decision-makers hold differing ideas, resulting in slower progress in the production line. Other issues that may affect the decision-making process include work and time pressure, planning weekend work shifts, decisions with partial fairness, and budget constraints, which can lead to project cancellations.
In light of the mentioned situations and challenges, certain qualities and characteristics define the decision-making process within the SME. These include carefulness in decision-making and, notably, consideration of risks, and subsequently accepting their responsibility. In addition to utilizing the skills matrix, which outlines employee task assignments, supervisors should also take into account shift scheduling and workload when making decisions. Considering the priority of tasks is also crucial, as it determines the alloca- tion of responsibilities, whether to highly qualified individuals or to those with lower experience. Based on the interviews, it was also found that most tasks in the company are repetitive, with rare occurrences of new tasks. This could explain why decision-making relies more on experiential knowledge, prompting employees with less experience to frequently seek guidance from their supervisors to ensure the accuracy of their decisions. While decisions are primarily reached collectively through discussions and regular meetings, hierarchical decision-making, particularly at the managerial level, is also observed.
Involvement Involving employees in decision-making not only significantly enhance orga- nizational efficiency but also fosters their creativity and commitment (Charles et al., 2021). Consequently, involvement is evaluated based on three key components: the opportunities for employee’s objection and suggestion, the methods of transparency regarding decisions made, and the importance of trust among team members; which are interrelated concepts according to research findings (Rawlins, 2008, Grates, 2007, Chene and Chr, 2011). The interviewees highlighted that suggestions and objections receive primary con- sideration when they contribute to enhancing production or when failures or problems are identified in the production process. Additionally, while the ideas are conveyed to the management team through supervisors, annual employee discussions also provide an opportunity for individuals to express their views directly. Moreover, opinions regarding staff assignment planning can be widely acknowledged. Surprisingly, some employees may refrain from
118 Gheibi and Böschen
offering suggestions or objections, arguing that intervening in superiors’ deci- sions is not expedient, and that everyone should focus solely on their own work, believing this approach leads to a stress-free environment.
The responses to transparency questions can be divided into two cate- gories: On one hand, decision-making transparency can be achieved through explanations from experts, addressing uncertainties raised by employees, clarifying new themes and decisions during weekly and monthly meetings, involving all relevant staff in decision-making, and fostering expertise and knowledge among the team over time. On the other hand, the transparency of decisions is context-dependent, and in certain cases, some decisions, particularly those at managerial levels, may not require understanding by others.
The participants emphasized the high level of trust among the team, high- lighting that without trust in their team members, progressing in the work process would be impossible. They remarked that this existing trust mini- mizes the need for negotiation on every subject, thereby accelerating the work process. Direct interaction between team members was cited as a key factor in increasing trust, leading employees to have complete trust in their superiors due to awareness of their knowledge and experiences. Additionally, system functioning contributes to building trust between superiors and employees. However, mismatches between decisions reached and their implementation in the workplace were noted to harm the trust employees have in their superiors.
Expectations An essential purpose of this study is exploring optimization prospects for a decision support system while considering the experiential insights of employ- ees engaged in the workflow. To achieve this goal, the third dimension investigates participants’ expectations of this system with the aim of sim- plifying, accelerating and ensuring the work process. Given that the primary objective of a decision support system is to simplify workflow by offering solutions, it is crucial for interviewees to have access to multiple proposed solutions for various scenarios. Recognizing work and time pressures as key challenges in decision-making, employees express a preference for reduced human intervention and alleviated time pressure through the digitalization of time-consuming and effortful tasks; For instance, the systematic catego- rization of the severity of various incoming errors would allow employees to easily identify the appropriate course of action when encountering spe- cific types of errors. This approach would bypass the necessity for extensive time allocation to initially determine the error type, thereby reducing uncer- tainty in decision-making processes. Additionally, there is a desire for data within the system to be organized in a comprehensible manner, especially considering the vast amount of data across the entire company. Providing self- familiarization tools for both new hires and existing staff with the workflow is also seen as essential for facilitating the work process. Another highlighted point focuses on relieving employees and reducing workload by delegating risk assessment tasks to the system. Within this context, decision-makers
Democratization in Industry via Multi-Agent Systems? The Case of a Production Company 119
also express a desire to automate fault attribution processes, potentially by establishing specific criteria to streamline the determination of responsibil- ity for particular damages or errors, thereby simplifying decision-making procedures.
The expressed expectations encompass points aimed at accelerating the work process. Recognizing staff assignment planning as a particularly time- consuming task, especially on weekends, underscores the necessity of imple- menting an automated system for assignment planning to assist supervisors, thereby significantly enhancing workflow efficiency and alleviating work- related pressure. Moreover, considering the aforementioned work pressure within the company and the potential for oversight regarding certain tasks and deadlines, employees have proposed the implementation of a signalling reminder system for upcoming plans. Additionally, access to a well-defined procedural framework would reduce challenges in finding experts and super- visors for consultations, thus furthering the acceleration of the work process. Other strategies, such as personally managing vacations and sick leave arrangements, along with task assignments tailored to individual preferences and competencies, serve to increase workflow efficiency. In addition, the implementation of automated customer service functionalities, such as email drafting, has the potential to mitigate time pressure. From a more technical perspective, it is advantageous to establish a structured inspection sequence for prioritizing error, enabling employees to accelerate the identification of the primary issue and address other errors effectively. The fulfilment of these requisites not only fosters punctuality but also enhances staff productivity and creativity.
The expectations articulated by the 8 interviewees regarding their antici- pated functionalities of this system can significantly contribute to the decision support system’s development and ensure the effectiveness of its workflow processes. The suggestions provided underline the critical need for accessing precise and reliable solutions, alongside secure access to all production and employee data. Therefore, establishing a comprehensive database contain- ing all relevant company information is imperative. Additionally, enabling data and workflow reviews can enhance the accuracy of system data and contribute to the efficiency of the decision-making process. Moreover, it is essential to provide insights into the consequences of proposed solu- tions, ensuring a thorough understanding of potential outcomes. Indicating resource availability and their precise whereabouts through system would ensure the progression of workflow processes, covering aspects such as product, machinery and even substitute employee availability. From another perspective, employees anticipate that the integration of a decision support system would ensure equitable treatment for all staff members. However, the comprehensiveness of this system should not hinder critical thinking, as cre- ativity and innovation must remain essential. Finally, there is an emphasis on protecting human discretion, recognizing the value of human judgment and autonomy within decision-making frameworks. Particularly, the participants envision the system as a tool that provides guidance, with ultimate decisions resting in human hands.
120 Gheibi and Böschen
CONCLUSION
This paper undertakes an exploration of industrial democratization through Multi-Agent Systems, focusing specifically on a case study of a production company. Through the utilization of a comprehensive methodology involv- ing documentary analysis, observations, and interviews, we have acquired valuable insights into the current decision-making processes, the degree of employee engagement within SMEs, as well as the anticipated functionali- ties of a decision support system by employees. The findings underscore the need for the development of a robust decision support system to initially offer production-line responsible persons a diverse array of solutions, thereby facil- itating informed decision-making. Subsequently, such a system can serve as a tool to provide relevant data, whether related to workforce assignment plans, customer data, inventory management, or other important factors such as components, employees, and timestamps. Providing data, solutions, and the experienced consequences via a system allows to asses superiors’ experience and simplifies the familiarization process for each newly hired employee. Moreover, they get enabled to gain sufficient confidence to make informed decisions. The implementation of such a system, coupled with the delegation of tasks and responsibilities to the system, shows potential for alle- viating existing workloads and time pressures, ultimately fostering employee productivity and innovation across all organizational levels.
With regard to the exploration of employee involvement, we found that the dominant high level of trust among team members has led to a decreased ten- dency to raise objections or offer suggestions. However, there is a desire for self-management of personal time-off (PTO) and task assignments through the system. Despite the presence of a strong level of trust within the team, such trust may be weakened in instances where a mismatch arises between deci- sions reached collectively and their subsequent implementation within the company. However, with the implementation of a decision support system, such problems can be addressed, as they may arise from the involvement of several decision-makers, each with potentially conflicting perspectives and suggestions. In addition, the analysis of outcomes within the involvement dimension indicates that decisions made at the shop floor level are typically characterized by a higher degree of transparency, often resulting from regular meetings, active inquiry, and the inclusive participation of all relevant staff members. Conversely, decisions formulated at the upper management level are less frequently perceived as transparent to the entire staff. Furthermore, notably, employees view the decision support system as a supplementary tool, rather than a replacement for human decision-making.
However, there is a tension emerging. The tension between the needs and ambitions of the ones being represented in the decision support tool and the forms and logics of representing within the tool. This tension can be seen as the litmus test of the development of these kinds of tools. If there are no further options to get an insight or to intervene into the ways of being repre- sented, the tool’s legitimacy is at risk. In alignment to this, the workers express a firm belief that such a system should not take away their autonomy in decision-making but should instead enhance their ability to make informed,
Democratization in Industry via Multi-Agent Systems? The Case of a Production Company 121
efficient, and simplified decisions. The consensus among employees is that the decision support system should act as a guiding mechanism, providing insights and recommendations to facilitate improved decision-making pro- cesses, while the ultimate responsibility for decision-making remains with human actors.
ACKNOWLEDGMENT
This work has been supported by the FAIRWork project (www.fairwork- project.eu) and has been funded within the European Commission’s Horizon Europe Programme under contract number 101069499. This paper expresses the opinions of the authors and not necessarily those of the European Com- mission. The European Commission is not liable for any use that may be made of the information contained in this paper.
Thank you to everyone who played a part in this project. Followed by. Johannes Zysk, Johanna Werz, Lucas Paletta, Jochen Mosbacher, Roland Sitar, Martin Pecnik, Julia Tschuden, Herwig Zeiner, Sylwia Olbrych, Gustavo Vieira.
REFERENCES
Bachrach, P. J. 1967. The theory of democratic elitism: A critique. Butzlaff, F. J. P. S. R. 2023. Consenting participation? How demands for citizen participation and expert-led decision-making are reconciled in local democracy. 21, 340–356. Charles, J., Francis, F. & Zirra, C. J. A. O. B. R. 2021. Effect of employee involvement
in decision making and organization productivity. 9, 28–34. Charles, L., Xia, S. & Coutts, A. P. J. I. L. O. R. 2022. Digitalization and Employment.
1–53. Chene, M. J. T. I. & Chr, M. I. 2011. Good practice in strengthening transparency, participation, accountability and integrity. Dingwerth, K., Schmidtke, H. & Weise, T. J. E. J. O. I. R. 2020. The rise of democratic
legitimation: why international organizations speak the language of democracy. 26, 714–741. Grates, G. 2007. Edelman Trust Barometer 2007. Edelman. Hilton, S. K., Arkorful, H. & Martins, A. J. M. R. R. 2021. Democratic leadership and organizational performance: the moderating effect of contingent reward. 44, 1042–1058. Nabatchi, T. 2012. A manager’s guide to evaluating citizen participation, IBM Center
for the Business of Government Washington, DC. Pantazi, M., Papaioannou, K. & van Prooijen, J. W. J. P. P. 2022. Power to the people: The hidden link between support for direct democracy and belief in conspiracy theories. 43, 529–548. Rajagopal, N. K., Qureshi, N. I., Durga, S., Ramirez Asis, E. H., Huerta Soto, R. M.,
Gupta, S. K. & Deepak, S. J. C. 2022. Future of business culture: an artificial intelligence-driven digital framework for organization decision-making process. 2022, 1–14. Rawlins, B. R. 2008. Measuring the relationship between organizational transparency and employee trust. Reuten, G. 2021. The employment performance of the Mondragon worker cooperatives 1983–2019.
122 Gheibi and Böschen
Stratu-Strelet, D., Gil-gómez, H., Oltra-Badenes, R. & Oltra-Gutierrez, J. V. J. J. O. B.
R. 2023. Developing a theory of full democratic consolidation: Exploring the links between democracy and digital transformation in developing eastern European countries. 157, 113543. Yin, R. K. 2014. Case study research: Design and methods (applied social research methods), Sage publications Thousand Oaks, CA.