Sustainable Lean Performance Potential Amidst the Transition Process from Industry 4.0 to Industry 5.0

Abstract

This study examines how selected technological and human-centered factors affect sustainable lean performance potential (SLPP) in manufacturing enterprises within the context of Industry 4.0 transition to Industry 5.0. The relationship between SLPP and Industry 4. transition to Industry 5.0 is contextual, meaning this direct relationship is not analyzed via statistical methods. A structured survey was conducted with 128 managers (n = 128), focusing on human-centric technology design (HCTD), artificial intelligence for waste minimization (AIWM), predictive maintenance (PMAI), and IoT integration in production (IOTP). The data were analyzed using descriptive statistics, correlation analysis, linear regression analysis, Harman’s single-factor test results, multicollinearity test, and non-linear curve estimation analysis. The results show that all four independent variables are positively associated with SLPP. IoT integration and AI for waste minimization had the strongest effects, followed by predictive maintenance. Human-centric technology design showed a weaker, yet still statistically significant, relationship with SLPP. The findings support a model where digital tools contribute directly to lean and sustainable outcomes, while human-centered approaches are emerging more gradually. The research adds empirical evidence to ongoing discussions about factors affecting lean performance in the context of industrial changes.

1. Introduction

Sustainable lean performance potential (SLPP) reflects the ability of enterprises to maintain efficiency, reduce waste, and remain adaptable over time [1,2]. While traditional models focused on standard lean principles, recent technological and organizational changes have introduced new ways to support these goals. The evolution of industrial systems, often discussed through the transition from Industry 4.0 to Industry 5.0, provides a useful background for understanding the changing conditions under which SLPP is developed [3]. Industry 4.0 introduced tools such as cyber-physical systems, the Internet of Things (IoT), and automation, which improved speed and consistency in production processes [4,5,6,7,8,9,10]. These technologies allowed enterprises to collect large amounts of operational data and adjust processes based on real-time information. In parallel, Industry 5.0 emerged with increased attention on human involvement, sustainability, and ethics [4,5,6]. This change highlighted both technical advancements and organizational practices that encourage employee participation, ensure safety, and promote long-term system stability [11,12]. Technologies such as IoT have become relevant to SLPP due to their ability to improve transparency and responsiveness in operations [13,14]. Real-time monitoring supports faster decisions, more efficient resource use, and adjustments to meet both performance and worker needs [15,16,17,18]. In manufacturing environments, this leads to better alignment between operational goals and human comfort, which can increase both productivity and satisfaction [19,20]. Artificial intelligence also supports SLPP, particularly in predictive maintenance and resource management [21,22,23]. Analyzing sensor data helps detect failures early, reduce downtime, and lower maintenance costs [24,25]. Combining AI and IoT supports systems that are stable, adaptable, and more capable of maintaining lean practices over time [26,27,28]. These advancements indicate that achieving sustainable lean results depends not just on the tools themselves, but also on how technologies are used to support both individuals and operational processes. This understanding reflects the direction of Industry 5.0, which sees humans and machines working in coordination to improve operational continuity and reduce waste [29,30,31,32,33].

Based on this, the transition toward Industry 5.0 introduces an additional dimension to SLPP. Rather than focusing only on technical performance, the new approach includes the human aspect, system resilience, and social sustainability. SLPP in this context is not limited to operational metrics but also reflects how well an enterprise supports employee well-being, interaction with smart systems, and adaptability in uncertain conditions. Human-centric technology design, responsible AI use, and ethical integration of automation are increasingly seen as necessary conditions for sustaining lean outcomes in dynamic environments. This shift from a purely technology-centered model to one that combines digital and human-centered values expands the meaning of lean sustainability. The transition does not replace Industry 4.0 foundations but adds new priorities that influence how efficiency and adaptability are achieved. SLPP serves as a way to assess how well enterprises are managing this transition. It connects earlier improvements based on digital tools with emerging expectations for inclusion, long-term usability, and socially aware operations. Therefore, the relationship between SLPP and Industry 4. to Industry 5.0 transition is contextual in nature. SLPP does not have a direct relationship, but rather the transition is viewed as framework within which the study is conducted.

The novelty of this study lies in its empirical examination of how selected variables within the context of the Industry 4.0 transition to Industry 5.0 affect sustainable lean performance potential (SLPP). The observed variables are human-centric technology design (HCTD), AI for waste minimization (AIWM), predictive maintenance (PMAI), and IoT integration in production (IOTP). While only HCTD is directly framed as a human-centric construct, the remaining variables contribute indirectly to human-centered design through their impact on safety, well-being, and operational transparency. For instance, IoT supports real-time monitoring of working conditions [34,35], AI reduces the mental workload by streamlining decisions [36,37], and predictive maintenance improves equipment reliability and safety [38,39].

This research positions itself within the transition process from Industry 4.0 to Industry 5.0 by examining how enterprises are integrating both advanced technologies and emerging human-centered priorities. The selected variables reflect this transitional phase, where traditional efficiency goals are complemented by sustainability and human-centric considerations. Although AI and IoT were central to Industry 4.0, their connection to waste reduction, predictive methods, and employee interaction in this study shows how their role is changing within a human-centered approach and the goal of sustainable lean performance.

The paper consists of six main sections. First, the Introduction provides a brief overview of the main subject. Next, the research background is presented. This includes the knowledge gap and hypotheses. The third section presents the Methodology and research phases. Afterwards, the results of the statistical analysis are noted. In addition, in this section, the developed theoretical framework is presented. Then, in the fifth section the obtained results are discussed, limitations are addressed, implications are noted, and future research ideas are given. Finally, the conclusions are based on the overall study.

2. Research Background

2.1. From Industry 4.0 to Industry 5.0

With the introduction of Industry 4.0, technologies such as cyber-physical systems and the Internet of Things allowed for greater automation, data collection, and machine-to-machine communication [40,41]. These tools supported lean strategies by enabling faster responses and improved process control [40,42,43]. However, most Industry 4.0 systems reduced the involvement of human input, placing greater emphasis on achieving speed and uniformity through automation [44,45]. This often raised concerns about job reduction and loss of workplace knowledge [46,47,48].

In response to these limitations, Industry 5.0 introduced a shift that brings human input back into focus. Rather than relying exclusively on automation, current practices explore how workers and machines can operate together in production environments [7]. The application of artificial intelligence, collaborative robots, and augmented reality has supported this shift, making it possible to distribute tasks based on both human and machine strengths [49,50,51]. This combination allows for leaner processes that adapt more easily to variable production conditions while preserving the role of human decision-making [52]. Personalization is an important element in developing lean strategies. Rather than depending on uniform mass production, many companies are adopting digital technologies and modular systems to provide customized solutions that align with specific customer needs [53]. These approaches must still comply with lean principles, such as reducing waste and maintaining short lead times. When properly supported by trained personnel and suitable technologies, personalization can be compatible with stable and efficient lean performance [54,55,56].

SLPP is also shaped by the need for operational resilience. Frequent disruptions, including global events, supply chain instability, and rapid market shifts, require systems that can adjust quickly and continue functioning under changing conditions [57,58]. Human–machine collaboration supports this by creating processes that can respond with greater flexibility. Resilient systems are more likely to sustain lean outcomes even during periods of instability, contributing to long-term operational continuity and reduced variability in performance [59]. The inclusion of human-centered design in industrial settings is another key aspect. Sustainable lean performance potential relies not just on efficiency, but also on creating systems that protect worker well-being and enable safe use of advanced technologies [26]. Human-centric systems use intuitive interfaces and provide a structure in which people can contribute actively. To sustain this, workforce development is necessary. Programs that strengthen digital literacy, collaborative thinking, and problem-solving help maintain performance as industrial settings evolve [60,61,62].

These developments suggest that SLPP is not shaped by technology alone, but also by how technologies are used in ways that support people and organizational goals. The broader transition from Industry 4.0 to Industry 5.0 reflects a gradual shift toward approaches that prioritize adaptability, social responsibility, and stable lean outcomes over time [14].

2.2. Human-Centric Approach and the Importance of IoT

Sustainable lean performance potential (SLPP) relies on systems that ensure operational efficiency while also addressing the work environment and user interaction. A key element influencing this is the design of systems centered human needs. When systems are easy to use and supportive rather than complex, workers and users can interact with them more effectively [34,63]. This can reduce stress and improve task completion, leading to better outcomes in production and service environments [15]. Designing systems with human needs in mind requires attention not only to technical aspects, but also to emotional and cognitive factors such as safety, comfort, and long-term usability [64,65]. The Internet of Things (IoT) supports SLPP through its role in human-centric operations. IoT, as a system of interconnected devices that collect and share data [17], allows environments to respond to real-time conditions. When applied with care, IoT can provide services in a way that avoids adding unnecessary complexity for the user [60,66,67]. These devices can operate continuously in the background, helping maintain safe and efficient workflows while requiring minimal manual input [50]. In work environments, sensors can monitor conditions like lighting, air quality, and noise to maintain comfort and reduce health risks [68,69,70]. Wearable IoT technology can also support worker well-being by tracking fatigue or exertion and providing feedback [61]. In production systems, IoT contributes to SLPP by reducing unplanned downtime, identifying machine issues early, and supporting safe operations [71,72].

Maintaining SLPP with the use of human-centered technologies also involves tackling challenges related to data privacy, system security, and accessibility. Since IoT devices often collect personal data, questions arise about who has access to this information and how it is safeguarded [30,73]. Security features must be built into systems from the start rather than added as secondary components [25,27]. In addition, inclusive design is necessary to ensure that SLPP can be maintained across diverse workforces. Different levels of digital literacy, access to technology, and personal needs must be accounted for [62,63,64]. Systems that can adapt to different users are more likely to support consistent and equitable lean outcomes.

The use of IoT within human-centric frameworks reflects a broader shift in how industrial systems are designed. The emphasis is no longer only on performance metrics but also on how systems contribute to satisfaction, participation, and steady work output [13]. Smart technologies are increasingly viewed not just as tools, but as parts of integrated environments that support people as they work and grow [63].

2.3. AI Application and Predictive Maintenance

As manufacturing environments adopt smarter technologies, predictive maintenance has become a practical way to maintain stable and efficient operations. It supports sustainable lean performance by reducing downtime, improving maintenance planning, and preventing premature equipment replacement [24,25]. This contributes to resource efficiency and process stability, both of which are consistent with lean objectives [23,26].

The introduction of artificial intelligence (AI) into industrial maintenance has changed how enterprises monitor and manage their systems. In place of rigid schedules or reactive repairs, AI makes it possible to predict when a machine is likely to require service [74,75,76]. This shift allows maintenance to occur before a failure happens, reducing interruptions and unnecessary interventions [77]. Modern machines are equipped with sensors that generate large volumes of operational data, such as temperature, vibration, or noise levels [78,79]. AI algorithms analyze these data streams to detect subtle patterns and early warning signs that might be missed by human operators [80,81]. With sufficient historical data, these systems can estimate when components may fail and help plan maintenance accordingly [74]. AI methods are especially valuable when working with systems that involve many variables or complex interactions. Traditional approaches often fall short in such settings [24]. In comparison, machine learning models are capable of detecting patterns among various signals and become more accurate as they are exposed to more data over time [82,83]. This continuous learning allows maintenance decisions to be more accurate, which helps maintain performance while keeping resource use efficient [11]. Predictive maintenance powered by artificial intelligence provides various operational benefits that support lean and sustainable performance. It allows for maintenance to be planned in advance, helping to prevent sudden breakdowns and avoid the interruptions, delays, and expenses caused by unexpected equipment issues [56]. Instead of replacing parts at fixed intervals, which may result in discarding functional components, decisions are based on actual signs of wear or system behavior. This approach extends the useful life of machinery and helps reduce unnecessary maintenance expenses [27].

A range of AI techniques is used to support predictive maintenance. Supervised learning is often applied, where models are trained on labeled datasets that include past instances of both normal operation and system failure [84]. These models can then process new data and estimate the likelihood of future failure. In situations where labeled data are limited, unsupervised learning can detect patterns or anomalies without predefined classifications [49]. This method is useful when identifying deviations that may indicate a developing issue, even if the cause is not yet fully known. Some enterprises also apply deep learning, using multi-layered neural networks to process complex inputs like audio signals or visual data from inspection tools [33].

There are also challenges in using AI for predictive maintenance. The accuracy of predictions depends heavily on the quality of the data. Inconsistent or incomplete sensor readings can reduce model reliability and lead to misleading conclusions [30]. A reliable data infrastructure is essential to make sure the input fed into models accurately represents real operating conditions. Another important issue is the clarity of AI-generated results. Some models may yield precise outcomes but lack understandable explanations, making them less practical for teams that need transparency in their decision-making process [35]. Research is ongoing to improve model explainability without reducing performance [15].

The shift toward predictive maintenance reflects a broader move from reactive responses to data-informed operational planning [3]. As AI techniques continue to advance and sensor integration becomes more common, this approach is expected to be widely adopted. Enterprises that implement predictive maintenance are more likely to achieve consistent performance, better resource management, and safer working conditions [77,83].

2.4. Knowledge Gap and Hypotheses

The current body of research on the transition from Industry 4.0 to Industry 5.0 has largely focused on technological advancements, efficiency improvements, and digital integration [6,9,21,85,86,87,88,89]. While these topics are widely studied, there is still limited exploration of how human-centric design, artificial intelligence for sustainability, predictive maintenance, and IoT integration affect lean performance in a real-world industrial context. Many studies discuss these elements separately, often within theoretical or conceptual frameworks, but empirical studies that combine them in a single model are relatively few. This creates a gap in understanding how these key elements interact and jointly influence sustainable lean performance potential. Another important gap is the lack of practical, survey-based studies that measure the perceptions and practices of enterprises related to these technologies. This makes it difficult for enterprises to prioritize efforts or investments when transitioning toward more human-centered and sustainable operations. Gaining a better understanding of how these technologies influence sustainable lean performance potential is important for making more informed operational decisions. Although digital tools provide advantages, their real value comes from supporting consistent, flexible, and efficient operations in different conditions. SLPP represents a broad indicator that captures both process efficiency and the lasting success of lean efforts in settings shaped by technological advances and a focus on human needs.

This study addresses these gaps by focusing on how selected Industry 5.0-related elements support SLPP. The integration of human-centric technologies, when assessed through the lens of lean sustainability, offers a more realistic evaluation of technological value beyond efficiency alone. Identifying these relationships helps guide the alignment of strategic goals, workforce development, and technological investment in ways that support both performance and long-term stability.

Through survey data collected from enterprises and analyzed with regression and correlation techniques, the study provides empirical evidence that supports or rejects the proposed relationships. This contributes to a better understanding of how Industry 5.0-related concepts can be linked to lean performance, offering new insights for both academic researchers and industry practitioners.

Based on the literature review and the main goal of the study, the following hypotheses are noted:

H₁. 

Human-centric technology design (HCTD) positively affects sustainable lean performance potential (SLPP).

H₂. 

AI for waste minimization (AIWM) positively affects sustainable lean performance potential (SLPP).

H₃. 

Predictive maintenance (PMAI) positively affects sustainable lean performance potential (SLPP).

H₄. 

IoT integration in production (IOTP) positively affects sustainable lean performance potential (SLPP).

While the formulated hypotheses are grounded in literature that highlights the relevance of each construct within Industry 5.0, the assumption of their universal applicability may not hold across all industrial contexts. The current study tests these relationships in a sample of enterprises with varying levels of technological maturity in a transitional economy. Therefore, the results reflect the influence of these factors under conditions where both early adopters and slower-moving enterprises coexist. Future research is encouraged to validate these hypotheses in more advanced and more digitally mature settings.

3. Methodology

To explore the hypotheses and provide insight into the effects of the independent variables (HCTD, AIWM, PMAI, and IOTP) on the dependent variable SLLP. The study applies a structured methodological process based on established variables from the literature. This methodology section outlines the logical steps taken to gather, prepare, and examine the data in a systematic way. The following subsection describes the research framework, including the definition of variables, structure of the data set, and details about the survey and analysis methods. This allows for a clear understanding of how the model was constructed and how the results were obtained.

3.1. Research Framework and Data Set

The research framework was designed to examine the relationship between selected Industry 5.0 independent variables (HCTD, AIWM, PMAI, IOTP) and the sustainable lean performance potential (SLPP) of manufacturing enterprises. To operationalize the independent variables, each construct was measured using multiple items in a 7-point Likert-scale format based on prior validated instruments and adapted to the industrial context. For Human-Centric Technology Design (HCTD), items assessed the presence of user-friendly interfaces, worker safety considerations, and training practices. AI for Waste Minimization (AIWM) included items related to the use of AI for resource efficiency and sustainability. Predictive Maintenance (PMAI) was measured through items reflecting the use of sensor data and predictive scheduling. IoT Integration in Production (IOTP) included items assessing the use of real-time monitoring and connected devices for operational control. These items are fully documented in Appendix A,

Table A1. Survey, attributes, available answers, and predictor groups.

Demographic Items [8]
Variable	Available Answers
Gender	○ Male ○ Female
Age	Individual values, range was not given
Education (acquired)	○ High school ○ Bachelor’s degree ○ Master’s degree ○ PhD
Enterprise size	○ micro (<10 employees) ○ small (10-49 employees) ○ medium (50-249 employees) ○ large (250> employees)
Main industry within the enterprise conducts business	○ manufacturing ○ textile ○ agriculture ○ mining and quarrying ○ information and communication ○ wholesale and retail ○ construction ○ finance and insurance ○ healthcare and social work ○ water, electricity, gas supply ○ education ○ other
Human-centric technology design (HCTD) [3,4,13,19,32]
7-point Likert-scale items 1 means “Totally disagree with statement”, 7 means—“Totally agree with statement”
Item	Available answers
1. Our technology implementations prioritize improving worker well-being and safety.	1   2   3   4   5   6   7 Totally disagree               Totally agree
2. We provide comprehensive training to ensure employees can effectively interact with new technologies.	1   2   3   4   5   6   7 Totally disagree               Totally agree
3. Feedback from employees is regularly sought to improve technological tools and systems.	1   2   3   4   5   6   7 Totally disagree               Totally agree
4. We assess the impact of new technologies on job satisfaction and employee morale.	1   2   3   4   5   6   7 Totally disagree               Totally agree
5. User-friendly interfaces are a key consideration in our technology adoption decisions.	1   2   3   4   5   6   7 Totally disagree               Totally agree
6. Our technological systems are designed to augment human capabilities rather than replace them.	1   2   3   4   5   6   7 Totally disagree               Totally agree
AI for waste minimization (AIWM) [33,57,70]
7-point Likert-scale items 1 means “Totally disagree with statement”, 7 means—“Totally agree with statement”
Item	Available answers
1. We use AI-driven analytics to identify and eliminate production inefficiencies.	1   2   3   4   5   6   7 Totally disagree               Totally agree
2. Continuous improvement initiatives are supported by AI-driven data analysis in our organization.	1   2   3   4   5   6   7 Totally disagree               Totally agree
3. We use AI to monitor and reduce emissions and environmental impact.	1   2   3   4   5   6   7 Totally disagree               Totally agree
4. Our organization invests in AI research and development focused on sustainability.	1   2   3   4   5   6   7 Totally disagree               Totally agree
5. AI technologies could improve our ability to achieve lean manufacturing goals.	1   2   3   4   5   6   7 Totally disagree               Totally agree
Predictive maintenance (PMAI) [23,25,38,69]
7-point Likert-scale items 1 means “Totally disagree with statement”, 7 means—“Totally agree with statement”
Item	Available answers
1. We employ predictive maintenance techniques to foresee and prevent equipment failures.	1   2   3   4   5   6   7 Totally disagree               Totally agree
2. Our maintenance schedules are optimized based on predictive analytics.	1   2   3   4   5   6   7 Totally disagree               Totally agree
3. Predictive maintenance contributes to our overall production efficiency.	1   2   3   4   5   6   7 Totally disagree               Totally agree
4. We collect and analyze equipment performance data to inform maintenance activities.	1   2   3   4   5   6   7 Totally disagree               Totally agree
5. We use predictive maintenance to extend the lifespan of critical machinery.	1   2   3   4   5   6   7 Totally disagree               Totally agree
6. Employees are trained to interpret and act on predictive maintenance data.	1   2   3   4   5   6   7 Totally disagree               Totally agree
IoT integration in production (IOTP) [22,75,97]
7-point Likert-scale items 1 means “Totally disagree with statement”, 7 means—“Totally agree with statement”
Item	Available answers
1. IoT technologies could improve our ability to track and manage inventory levels in real-time.	1   2   3   4   5   6   7 Totally disagree               Totally agree
2. IoT systems could provide actionable insights that contribute to our lean objectives.	1   2   3   4   5   6   7 Totally disagree               Totally agree
3. IoT-enabled connectivity supports seamless information flow across departments.	1   2   3   4   5   6   7 Totally disagree               Totally agree
4. IoT integration can help reduce downtime and improve equipment utilization.	1   2   3   4   5   6   7 Totally disagree               Totally agree
5. IoT implementation can increase transparency and traceability in production processes.	1   2   3   4   5   6   7 Totally disagree               Totally agree

, allowing for replication and adaptation in future studies.

The selection of the four independent variables is based on the convergence of literature on sustainable manufacturing, digital transformation, and Industry 5.0. Human-centric technology design has been widely discussed as a foundational principle of Industry 5.0 [64,65,66], linking technological progress with worker well-being and engagement. AI for waste minimization reflects the growing body of research advocating AI’s role in resource efficiency, process optimization, and emission reduction [77,78]. Predictive maintenance is a practical application of AI and sensor technologies, aligned with lean principles focused on minimizing downtime and extending equipment life [80,81,82,83]. IoT integration in production has been identified as a core enabler of transparency, real-time control, and lean adaptability [67,68,69]. These variables were selected due to their frequent appearance in both theoretical discussions and applied research on the transition from Industry 4.0 to 5.0, and their direct relevance to sustainable lean outcomes.

The data for this study were collected using a structured survey instrument composed of 7-point Likert-scale items. This type of scale was selected because it allows respondents to express their opinions with a higher level of detail. The broader range of options helps capture subtle differences in perception and reduces the likelihood of respondents consistently choosing neutral answers. This improves the overall accuracy of the data and helps to reflect the actual views and practices of the participants more clearly. The questions were arranged in a randomized order across different sections of the survey to reduce the chances of patterned or automatic responses. To encourage honest participation, respondents were informed that their answers would remain anonymous. This also helped reduce any hesitation they may have felt in providing truthful responses.

To reduce bias in the way questions were interpreted, all survey items were written using neutral wording. In addition, the layout of the survey was structured so that predictor variables and outcome variables appeared in separate sections. This separation helped to reduce the influence of one set of questions on responses to another. These procedural precautions were taken to increase the reliability of the results and reduce the potential influence of common method bias.

The survey targeted managers working in manufacturing enterprises who were expected to have relevant knowledge about technology implementation and lean production practices. Invitations to participate in the survey were sent by email, using a curated list of contacts obtained through professional networks and industry-related databases. A total of 530 invitations were distributed, and 128 valid responses were received, resulting in a response rate of approximately 24.15 percent. The data collection phase lasted 3 months, during which follow-up emails were sent to encourage participation. Respondents were reminded of the voluntary nature of their involvement and were again assured that their responses would remain anonymous.

The final sample consisted of manufacturing enterprises from a variety of industrial sectors. These included metal processing, machinery, food and beverage, plastics, automotive components, and electronics. The enterprises also varied in size, with 7% classified as micro enterprises, 31% as small, 49% t as medium, and 13% as large. This diversity in both sector and size improved the representativeness of the sample. It also allowed for the inclusion of enterprises at different stages of digital transformation. Some enterprises reported already using certain technologies associated with Industry 4.0, such as IoT systems for monitoring production or AI tools for quality control. Others were still in the early stages of digital adoption, with efforts underway but not yet fully developed. This variation in technological maturity within the sample was valuable because it reflected the reality of many transitional economies, where some companies are more advanced while others are just beginning to adopt digital tools. As a result, the findings of this study offer insights into how lean performance potential is shaped under these mixed conditions.

To analyze the collected data, several statistical procedures were used. These included descriptive statistics, correlation analysis, and linear regression, which helped examine relationships between the key variables. Tests for multicollinearity were conducted to ensure that each independent variable contributed distinct information to the model. To explore the possibility of bias introduced through the data collection method, Harman’s single-factor test was carried out. This test helped determine whether one factor explained the majority of the variance, which could indicate a problem with common method bias. Additionally, a curve estimation analysis was performed to investigate whether non-linear relationships existed between the independent variables and the dependent variable, sustainable lean performance potential (SLPP). This step added further depth to the analysis and helped uncover patterns that may not have been evident through linear methods alone.

These methods were used to identify relationships between variables, assess the strength and direction of effects, and test the robustness of the model. The statistical significance and internal consistency of the constructs were also examined to validate the reliability of the findings. The hypotheses are presented within the research framework that is presented in Figure 1.

In

Table 1. Analyzed variables.

Variable	Information
Human-centric technology design (HCTD): Independent variable	Human-centric technology design (HCTD) focuses on designing and implementing technology in ways that support workers, rather than replace them. This includes prioritizing employee well-being, safety, and job satisfaction when introducing new systems. Human-centric design allows for more intuitive and adaptive technology, enabling workers to interact effectively with machines and benefit from the tools rather than being burdened by them [90,91].
AI for waste minimization (AIWM): Independent variable	AI for waste minimization (AIWM) refers to the use of artificial intelligence tools and algorithms to reduce material, energy, and time waste in production processes. AI can help detect inefficiencies, optimize resource use, and support continuous improvement efforts by analyzing data and suggesting better operational decisions [92,93].
Predictive maintenance (PMAI): Independent variable	Predictive maintenance (PMAI) involves using data from sensors and equipment to forecast when maintenance is needed. Instead of following a fixed schedule, predictive maintenance uses machine learning to identify early signs of wear or failure. This reduces unplanned downtime, extends equipment life, and improves production reliability. The variable reflects how organizations adopt this approach in practice [94].
IoT integration in production (IOTP): Independent variable	IoT integration in production (IOTP) measures how well connected devices and systems are within production environments. Internet of Things (IoT) technology helps track performance, inventory, and operational status in real time. Integration of IoT allows for more responsive, transparent, and efficient production processes, supporting lean and adaptive practices in manufacturing [95].
Sustainable lean performance potential (SLPP): Dependent variable	Sustainable lean performance potential (SLPP) represents the ability of an enterprise to achieve lean goals such as efficiency, waste reduction, and continuous improvement. The variable captures how enterprises combine lean principles with future-oriented industrial practices [96].

, short descriptions for each analyzed variable are presented.

In the

Table 2. Methodology summary.

Research Parameter	Information
Number of participants	128 (n = 128)
Research duration	From survey development to data collection: 3 months
Sample structure	manufacturing enterprises
Instrument for data collection	Survey with 7-point Likert-scale items (presented in Table A1)
Data analysis	descriptive statistics, correlation analysis, linear regression analysis

, the methodology summary is presented.

3.2. Conducted Research

First, a detailed survey was designed. The survey is presented in Appendix A, Table A1. The independent variables’ items and the dependent variable’s items are created based on the extensive literature analysis (references are noted within the survey).

The sampling method applied was purposive sampling, targeting managers with experience in digital transformation and lean production within manufacturing enterprises. There are approximately 15,900 manufacturing enterprises of which 69.7% apply some form of lean concept. This equals to 11,082 eligible enterprises, of which 8340 are micro, 1944 are small, 640 are medium, and 158 are big enterprises. According to the equation for finite populations:

n = [Z² ∗ p ∗ (1 − p)/e²] ∗ [N/(N + ((Z² ∗ p ∗ (1 − p))/e² − 1))]

(1)

where

Z: z-value for a given confidence level (for 95%, Z = 1.96).

p: population proportion of success. If unknown, maximum variability is assumed (p = 0.5), as it yields the largest required sample size.

e: acceptable margin of error (e.g., e = 0.05 for 5% error).

N: population size (e.g., total number of enterprises N = 11,082).

An error margin of 1% would require a significantly larger sample. A higher margin of error, usually between ±7% and ±10%, is a practical solution when resources are limited or when the research has less critical goals. In such cases, the sample is smaller and often used in preliminary studies, quick assessments, or when the population is relatively homogeneous. In this study, a standard error margin of ±5% will be used, along with a stratified sample based on enterprise size. For a total sample size of 38 enterprises, proportional to the size of each group, the sample distribution is as follows:

• Micro enterprises: 32
• Small enterprises: 4
• Medium enterprises: 1
• Large enterprises: 1

Enterprises were approached via personalized email invitations, which outlined the research purpose and confidentiality guarantees. The survey link was embedded in the email, and two follow-up reminders were issued during the 3-month data collection period. Sometimes, managers are reluctant to fill out forms; therefore, in order to acquire the minimum required sample of 38, we sent out a total of 530 invitations. A total of 128 complete responses were received, yielding a response rate of 24.15%. This response rate reflects a satisfactory level of engagement from the targeted population.

Next, the acquired data were stored on multiple locations (portable HDDs as well as on the cloud services) in order to ensure backups. The data were checked for missing values, inconsistencies, and outliers.

Afterwards, statistical analyses to evaluate the hypotheses and test the proposed model were conducted. This included descriptive statistics, correlation analysis, linear regression analysis, Harman’s single-factor test results, multicollinearity test, and non-linear curve estimation analysis.

This structured approach ensured methodological rigor, transparency, and replicability, providing a sound foundation for generating meaningful insights into the relationship between the independent variables and the dependent variable.

4. Results

The results of the descriptive statistics are presented in

Table 3. Descriptive statistics.

Dimension	Mean (μ)	Standard Deviation (σ)	Cronbach’s Alpha
Human-centric technology design (HCTD)	5.31	1.42	0.925
AI for waste minimization (AIWM)	5.22	1.40	0.912
Predictive maintenance (PMAI)	5.31	1.35	0.936
IoT integration in production (IOTP)	5.28	1.30	0.935
Sustainable lean performance potential (SLPP)	5.31	1.26	0.920

.

The results presented in Table 3 show high average scores across all variables, with means ranging from 5.22 to 5.31 on a 7-point scale. Additionally, all variables show strong internal consistency, with Cronbach’s alpha values above 0.91, confirming the reliability of the measurement instruments. Next, the results of the correlation analysis are presented in

Table 4. Results of the correlation analysis measurement instruments.

Correlation Analysis
	Human-Centric Technology Design (HCTD)	AI for Waste Minimization (AIWM)	Predictive Maintenance (PMAI)	IoT Integration in Production (IOTP)	Sustainable Lean Performance Potential (SLPP)
HCTD	1.000
AIWM	0.671 *	1.000
PMAI	0.339 *	0.423 *	1.000
IOTP	0.020	0.378 *	0.043	1.000
SLPP	0.148	0.366 *	0.338 *	0.452 *	1.000

* Significance 5%.

.

Table 4 provides insight into the relationships between the variables. All independent variables are significantly correlated with sustainable lean performance potential (SLPP), except for human-centric technology design (HCTD), which shows a weaker correlation. The strongest correlation with SLPP is observed for IoT integration in production (0.452), implying that IoT plays an important role in lean performance from the perspective of the respondents. Further, the results of the linear regression analysis are presented in

Table 5. Results of the linear regression analysis.

Regression Analysis
Y	X	β	p-Value	R2	F	F Sig.
		Intercept: 1.984
SLPP	HCTD	0.226	<0.0001	0.648	162.951	<0.0001
	AIWM	0.305	<0.0001
	PMAI	0.241	<0.0001
	IOTP	0.345	<0.0001
Y		X	MSE		RMSE	DW
SLPP		HCTD	0.204		0.468	2.065
		AIWM
		PMAI
		IOTP

.

The model for the linear regression with all variables has the following form: $SLPP=1.984+0.226·HCTD+0.305·AIWM+0.241·PMAI+0.345·IOTP+ϵ$.

All independent variables have statistically significant positive effects on SLPP, with p-values below 0.0001. The highest standardized coefficient (β = 0.345) is associated with IoT integration, followed by AI for waste minimization (β = 0.305), predictive maintenance (β = 0.241), and human-centric technology design (β = 0.226). This suggests that while all constructs contribute to lean performance, IoT has the most pronounced influence in the current sample. The R² value of 0.648 indicates that approximately 65% of the variation in SLPP is explained by the combined predictors, which demonstrates a strong model fit.

Common method bias arises when the measurement method used in a study, such as a single survey completed by the same respondents, introduces systematic variance that can distort the observed relationships between variables. This bias can inflate or deflate correlations between constructs, leading to misleading interpretations. To reduce this risk, several procedural steps were taken during the survey design. These included assuring respondent anonymity, separating the predictor and criterion variables in the survey layout, randomizing the order of questions, and avoiding suggestive wording. These steps were intended to lower the chances of response patterns, evaluation apprehension, and consistency motives influencing the answers.

In addition to these design measures, a statistical approach was used to assess whether common method bias was present in the dataset. Harman’s single-factor test was conducted by loading all survey items into an unrotated factor analysis. In

Table 6. Harman’s single-factor test results.

	Eigenvalue	Variance Explained (%)
Factor 1 (SLPP)	4.25	34.0
Factor 2 (HCTD)	1.85	14.8
Factor 3 (AIWM)	1.40	11.2
Factor 4 (PMAI)	1.30	8.8

, the Harman’s single-factor test results are presented for all four factors (SLPP, HCTD, AIWM, and PMAI).

The results show that no single factor dominated the variance structure, with the first factor accounting for less than 35% of the total variance. This is below the common threshold of 50%, suggesting that the variance observed in the data is distributed among multiple constructs rather than stemming from a single source. The emergence of several factors with eigenvalues above 1 also supports this interpretation, indicating that the items measured distinct concepts and that the risk of inflated relationships due to common method bias is minimal.

Furthermore, in

Table 7. Multicollinearity test.

	Human-Centric Technology Design (HCTD)	AI for Waste Minimization (AIWM)	Predictive Maintenance (PMAI)	IoT Integration in Production (IOTP)	Sustainable Lean Performance Potential (SLPP)
Tolerance	0.485	0.385	0.425	0.445	0.395
Variance Inflation Factor (VIF)	1.854	1.932	1.942	2.305	1.994

, the results of the multicollinearity test are presented.

With all Variance Inflation Factor (VIF) values below 2.5, it can be confirmed that multicollinearity is not a concern in this analysis and that each independent variable provides distinct explanatory power. Next, in

Table 8. Non-linear curve estimation analysis results.

Variable	Adjusted R-Squared (Polynominal)
HCTD	0.184
AIWM	0.225
PMAI	0.112
IOTP	0.251

the results of a non-linear curve estimation analysis are presented.

The curve estimation results, with adjusted R-squared values ranging from 0.112 to 0.251, indicate that non-linear effects between the independent variables and sustainable lean performance potential (SLPP) are present but modest. This aligns with the study’s context, which examines transitional manufacturing enterprises where adoption of Industry 5.0 elements varies in depth and maturity. IoT integration in production (IOTP) shows the strongest non-linear effect, suggesting that its influence on SLPP increases more rapidly at higher levels of implementation. This supports the main finding of the paper that IOTP is the most impactful predictor, particularly as more advanced IoT applications become embedded in operations. Similarly, AI for waste minimization (AIWM) shows a moderate non-linear relationship, indicating that its benefits may accumulate over time as systems evolve and improve, reinforcing its significant role in reducing inefficiencies.

Human-centric technology design (HCTD), while the weakest predictor in linear terms, shows a modest non-linear pattern. This suggests its full impact on SLPP may only become evident at higher levels of organizational alignment and cultural adaptation, which are slower to develop. Predictive maintenance (PMAI) displays the weakest non-linear effect, consistent with its steady, incremental contribution to lean outcomes, as reported in the linear analysis.

Overall, these results complement the original findings and suggest that while linear effects are dominant, some predictors, especially IOTP and AIWM, may yield increasing returns as adoption deepens. Recognizing these non-linear patterns can help managers better plan the sequencing and scaling of digital and human-centered innovations during the transition to Industry 5.0.

Furthermore, the obtained results helped in developing a theoretical model that includes the observed variables through the transition process from Industry 4.0 to Industry 5.0. The developed model is presented in Figure 2.

The theoretical model shown in Figure 2 presents the relationship between key enablers of Industry 5.0 and sustainable lean performance potential (SLPP). The model is structured around the four independent variables: human-centric technology design (HCTD), artificial intelligence for waste minimization (AIWM), predictive maintenance (PMAI), and Internet of Things integration in production (IOTP).

Each of these components is shown to have a direct and positive association with SLPP, based on the findings from the regression analysis. The model demonstrates how SLPP does not result from a single factor but from the combined influence of different technological and organizational developments. It also highlights how these developments reflect the ongoing shift in focus from purely efficiency-driven strategies to more balanced approaches that include social, ethical, and worker-centered considerations.

In the context of the transition between Industry 4.0 and Industry 5.0, SLPP can be understood as a measurable outcome that reflects how well a company is managing to maintain lean production principles. This includes waste reduction, reliability, and adaptability, workplace safety, inclusion, and long-term stability. During the earlier phase of digital transformation, represented by Industry 4.0, most changes were focused on automation, data collection, and system integration. These improvements typically targeted immediate productivity gains and were centered on tools such as IoT systems and automated equipment. In the model, this is reflected in the strong influence of IOTP and AIWM on SLPP, as confirmed by the statistical findings. These technologies are more frequently adopted because they bring visible improvements in performance and process efficiency. As organizations continue this transition, they begin to adopt practices more closely associated with Industry 5.0. These include systems that are easier to use, better aligned with human needs, and capable of supporting personalized workflows. Human-centric technology design reflects these priorities. Although it showed the weakest statistical effect in this study, it still demonstrated a positive contribution to SLPP. This suggests that while organizations are aware of the value of worker-centered systems, such systems may take longer to implement and produce effects that are less immediate but still relevant. Predictive maintenance also fits into this evolving picture, as it combines sensor-based monitoring with adaptive maintenance planning. Its consistent but moderate impact in the model indicates that enterprises use it to reduce unplanned downtime and keep operations steady over time.

The layout of Figure 2 visually supports the idea that sustainable lean performance is not limited to one domain of activity. The positioning of the four independent variables around the SLPP construct suggests that the transition from Industry 4.0 to Industry 5.0 is not linear. Instead, organizations progress at different speeds and in different ways depending on their resources, goals, and existing systems. Some may begin with data-driven tools, while others may prioritize worker training or maintenance strategies. This layered adoption process is reflected in the model, which shows SLPP as an integrated outcome of multiple actions rather than the product of a single strategy.

5. Discussion

5.1. Assessing the Obtained Results

The results highlight that certain technologies have a direct impact on sustainable lean performance potential. IoT integration and artificial intelligence for waste reduction were identified as the most influential, showing that companies are increasingly relying on connected systems, real-time information, and advanced analytics to optimize processes and limit waste. These technologies help operations react more quickly to changing needs while supporting lean principles. Predictive maintenance also showed a notable effect, pointing to a trend where companies are moving toward condition-based maintenance strategies. This approach helps prevent unexpected breakdowns and supports the long-term stability of processes needed to maintain lean performance. In contrast, the influence of human-centric technology design on SLPP, although statistically significant, was weaker than the other variables. This indicates that while the human-centered dimension is part of strategic conversations, it is not yet a strong driver of measurable lean outcomes in most enterprises. Practical implementation of worker-centered innovations appears to be in the early stages. Many organizations may still be focused on integrating digital technologies that deliver immediate operational results before undertaking deeper changes that reflect human-centric principles. The strongest effects observed in this study were linked to technologies that improve visibility and allow for direct resource optimization. This suggests that measurable gains continue to be a primary focus when enterprises adopt new solutions.

Rather than reflecting a complete shift, the results show a layered transition. Enterprises are building on existing digital infrastructure and using it to improve SLPP before fully integrating broader organizational values associated with Industry 5.0. Human-centric and resilience-oriented practices are being introduced gradually, often following initial success with more technical, efficiency-driven improvements. The findings suggest that the transition toward Industry 5.0 is being shaped by practical concerns tied to lean performance. Over time, as confidence and experience with digital tools increase, broader changes in culture and organizational priorities may follow.

Furthermore, on the research results, the four hypotheses are addressed:

• H₁: Human-centric technology design (HCTD) positively affects sustainable lean performance potential (SLPP). The regression results show a statistically significant positive effect (β = 0.226, p < 0.0001), indicating that HCTD does influence SLPP, although the effect is weaker compared to the other variables. Therefore, hypothesis H₁ failed to be rejected.
• H₂: AI for waste minimization (AIWM) positively affects sustainable lean performance potential (SLPP). The regression coefficient is positive and statistically significant (β = 0.305, p < 0.0001), showing that AIWM has a strong effect on SLPP. Thus, hypothesis H₂ failed to be rejected.
• H₃: Predictive maintenance (PMAI) positively affects sustainable lean performance potential (SLPP). The results indicate a significant positive relationship (β = 0.241, p < 0.0001), confirming PMAI as a relevant contributor to SLPP. Hence, this hypothesis also failed to be rejected.
• H₄: IoT integration in production (IOTP) positively affects sustainable lean performance potential (SLPP). It shows the strongest effect among all predictors (β = 0.345, p < 0.0001), highlighting IOTP as the most influential variable on SLPP in this study. Based on this, the hypothesis failed to be rejected.

All four hypotheses are supported by the empirical data and therefore failed to be rejected. The results show that each of the proposed Industry 5.0 variables significantly contributes to sustainable lean performance potential in manufacturing enterprises.

5.2. Assessing the Literature and Previous Findings

The results of this study align with many topics discussed in recent literature addressing the transition process from Industry 4.0 to Industry 5.0 [2,70,88,97]. A number of previous studies have noted the role of digital technologies as enablers of lean manufacturing and operational efficiency [10,28,56]. The strong statistical impact of IoT integration in this study is consistent with findings that describe IoT as a key driver of transparency, traceability, and real-time decision-making in manufacturing systems [18,20,98]. The ability of IoT to connect machines, collect data, and support automation aligns well with lean principles, especially in terms of minimizing waste and improving operations [19]. Similarly, the observed impact of AI for waste minimization supports earlier discussions that position artificial intelligence as a core component of smart manufacturing [40,46,68,76]. Previous studies often highlight AI’s capacity to process large datasets, identify inefficiencies, and guide decisions that lead to resource savings [29,43]. In this context, the significant contribution of AI to sustainable lean performance potential reinforces the idea that AI technologies are not only tools for automation but also strategic instruments for achieving sustainability and competitiveness [30,99].

Predictive maintenance has also been frequently presented in the literature as a practical application of AI and sensor technologies [69,81]. This aligns with the lean focus on maximizing equipment reliability and reducing non-value-added activities [75]. Compared to traditional maintenance approaches, predictive strategies are more responsive and data-driven, which explains their increasing popularity among enterprises aiming to stay agile and cost-efficient [41,73].

Another important observation relates to how strongly the combined model accounts for lean performance outcomes. With an R² value of 0.648, the model performs well and confirms that technological variables are more mature in application and perception than some of the broader human-centered elements, at least within the current sample [4]. While the literature often promotes an ideal balance between human and machine, actual practice may still favor technical implementation over holistic integration [32,57,58].

5.3. Limitations, Future Research, and Implications

This study has several limitations that should be considered when interpreting the results. First, the research is based on self-reported data collected through a structured survey. Although the survey items were carefully developed and tested for reliability, self-assessment may lead to bias, particularly in overestimating the extent of technological implementation or its perceived effectiveness. Future research could complement survey-based findings with case studies, expert interviews, or direct observations to gain a deeper understanding of actual practices. Second, the study was conducted within a specific national and industrial context, with respondents drawn exclusively from manufacturing enterprises operating in a single country. While the sample includes various sectors and enterprise sizes, the findings may not fully represent the conditions and technological maturity of other regions or industries. To strengthen generalizability, future research should consider comparative studies across countries or cross-sectoral analyses that capture different stages of Industry 5.0 adoption. Future research should also explore potential moderating or mediating variables that could explain the weaker correlation between human-centric technology design and lean performance. Factors such as organizational culture, digital maturity, leadership style, or employee engagement may influence how human-centric practices translate into operational outcomes. Investigating these interactions could provide understanding of how human and technological factors align in Industry 5.0 settings.

Another limitation affecting the validity of the results is the self-reported nature of the data. Respondents may have provided responses that reflect intended practices rather than actual implementation, particularly in areas involving newer or aspirational technologies such as human-centric system design. The subjective perception of SLPP and its enablers could be influenced by optimism bias or misunderstanding of terminology. Additionally, despite efforts to ensure sample diversity, these factors should be considered when interpreting the strength and direction of the reported relationships.

Although procedural and statistical steps were taken to reduce common method bias, its presence cannot be entirely ruled out. A Harman’s single-factor test was also performed, and results indicated that no single factor accounted for the majority of variance, suggesting that common method bias was not a major concern in this dataset.

Furthermore, the findings of this study have several social, political, economic, and managerial implications. Social implications are primarily connected to the role of human-centric technology design. Although the statistical impact of this variable was weaker than expected, its inclusion in the regression model confirms that it still holds relevance. This suggests that technological advancements alone are not sufficient for long-term progress unless they are paired with respect for human well-being, skills, and dignity. Socially, this means that enterprises must consider how automation and AI affect job roles, employee morale, and workplace inclusion. The adoption of technologies should not come at the cost of workforce displacement or dissatisfaction but should support learning, safety, and meaningful employment. A stronger commitment to social sustainability can build trust among employees and increase long-term organizational stability.

Political implications relate to policy development and regulatory frameworks that support the transition to Industry 5.0. Governments and public institutions play an important role in creating incentives, guidelines, and educational reforms that promote the integration of human-centered and sustainable practices in industry. The results suggest that technologies like IoT and AI are being adopted, but there is still a gap in the structured integration of human-centric principles. Policymakers can respond to this gap by developing standards and certifications. This way ethical technology use, protection of workers’ rights, and transparency requirements in AI and data systems would be implemented.

As for economic implications, the findings indicate that enterprises can achieve higher productivity, lower waste, and more agile operations by investing in these technologies. In the long term, such efficiency improvements contribute to competitiveness and economic resilience, especially in markets where resources are limited or supply chains are volatile. On a macroeconomic level, widespread adoption of these practices can support national productivity, reduce environmental externalities, and contribute to green growth agendas.

Managerial implications are directly connected to how enterprises structure their strategies and operational processes. Managers are encouraged to view Industry 5.0 not as a replacement for existing systems, but as an opportunity to align human and technological capabilities. The findings suggest that while investments in AI and IoT are likely to produce measurable improvements in lean performance, efforts should also be made to integrate human feedback, training, and workplace design into the technological change process. Managers should focus on building organizational cultures that value adaptability, data-driven decision-making, and cross-functional collaboration. Leadership must also prepare for new challenges related to data security, ethical decision-making, and interdisciplinary talent management.

6. Conclusions

The findings confirm that all four variables have a positive impact on SLPP, with IoT integration and AI for waste reduction standing out as the most influential. Predictive maintenance also made a notable contribution, while human-centric technology design, though statistically relevant, showed a weaker association. These results point to the increasing importance of smart, data-driven technologies in supporting lean and sustainable operations. The outcomes have practical value for enterprises working to improve lean performance within a framework that is both sustainable and centered on human needs. Additionally, the study provides empirical evidence linking elements of Industry 5.0 to lean manufacturing performance. The study contributes not only in terms of subject matter but also from a methodological perspective. Unlike many existing works that rely on conceptual discussions or technical assessments, this research applies a structured, survey-based approach to gather data directly from managers working in manufacturing enterprises. The use of empirical methods allowed for quantifiable testing of the relationships between the selected variables and SLPP. The inclusion of a 7-point Likert-scale survey, randomization of question order, and statistical validation procedures (such as Harman’s single-factor test and curve estimation) adds robustness to the findings. The model used in this research provides insight into how these variables interact within real industrial contexts, particularly in economies where digital transformation is uneven across enterprises.

From a findings perspective, the study shows that all four variables have a statistically significant positive effect on SLPP. IoT integration and AI for waste minimization showed the strongest effects, suggesting that these technologies are more established in practice and perceived as more directly tied to performance improvements. Predictive maintenance also showed a meaningful contribution, indicating that many enterprises use this strategy to reduce downtime and maintain consistent operations. Although human-centric technology design had the weakest effect, it was still statistically significant, which suggests that organizations are beginning to recognize its relevance even if implementation is less advanced. This combination of findings adds to the existing body of knowledge by showing how both human-centered and technical factors can jointly support sustainable lean outcomes during a period of industrial transition.

Future studies could apply comparative or multi-country analysis to account for cultural, regulatory, and economic differences that may influence the implementation and effectiveness of Industry 5.0 practices. This will allow enterprises, policymakers, and researchers to better understand and shape the future of manufacturing in line with Industry 5.0 principles.