Evolution and Key Differences in Maturity Models for Industrial Digital Transformation: Focus on Industry 4.0 and 5.0

Abstract

This study conducts an Academic Literature Analysis of 75 maturity models to clarify how Industry 4.0 and Industry 5.0 are being conceptualized and assessed. We map model scope, level structures, evaluated dimensions, and enabling technologies and complement descriptive statistics with exploratory non-parametric tests on the relationship between level structure and dimensional breadth. Results show a persistent dominance of Industry 4.0 models (≈92%), alongside a recent but steady emergence of Industry 5.0 and hybrid approaches in the latest models. Structurally, five-level schemes prevail, balancing diagnostic granularity and comparability. Content-wise, Technology and Digitalization, Processes and Operations, and Management and Strategy remain core, while People and Competencies and Innovation gain relevance; Sustainability and Social Responsibility and Human–Machine Interaction appear with the rise of Industry 5.0. We contribute (i) an operational definition of “hybrid” maturity models to make the I4.0→I5.0 transition measurable, (ii) a meta-typology of maturity levels explaining the five-level preference, and (iii) an evidence-based technology cartography across models. The findings suggest that future designs should retain the digital backbone of I4.0 while integrating explicit indicators for human-centricity, sustainability, and resilience with transparent weighting and scenario-based validation.

1. Introduction

The industrial evolution from Industry 1.0 to Industry 5.0 reflects a sequence of major shifts in how production is organized and controlled. The First Industrial Revolution replaced manual labor with mechanized processes powered by steam and water, dramatically increasing efficiency and output [1,2]. The Second Industrial Revolution built on this with electrification, mass production, and assembly lines, enabling large-scale, standardized manufacturing and spurring advances in the chemical and petroleum industries [1,2]. In the late 20th century, the Third Industrial Revolution introduced electronics, information technology, and automation, which improved process control and flexibility and marked the start of industrial digitalization [1,2].

Industry 4.0, often described as the Fourth Industrial Revolution, deepens this digitalization by integrating cyber–physical systems, the Internet of Things (IoT), artificial intelligence, big data analytics, and digital twins into production systems and business models [3,4,5,6]. These technologies enable automated and interconnected manufacturing, particularly in smart factories, where advanced sensors, real-time data, and IoT-based connectivity enhance performance, quality, transparency, and responsiveness to market signals [7,8,9]. Smart manufacturing systems require a shift in management philosophy and capabilities to exploit intelligent systems and remain competitive [10]. When adequately implemented, benefits include improved quality and safety, cost reductions, productivity gains, and competitive advantage over traditional production systems [11,12].

However, the transition to Industry 4.0 has also been complex and uneven. Organizations face managerial and organizational challenges, such as the need for strategic planning, new competencies, and changes in processes and structures [13,14]. Integrating cyber–physical systems, big data analytics, and other technologies into legacy environments is hindered by infrastructure constraints and financial limitations [13,15,16]. The perceived opportunities and challenges of Industry 4.0 vary by company size, sector, and whether the firm is a technology provider or user [14]. Local production systems, especially SMEs, may require place-based industrial policies to overcome multi-level technological gaps and fully exploit Industry 4.0 opportunities [17,18].

The current state of Industry 4.0 implementation also differs across countries and sectors. Some economies, such as Germany and China, have developed comprehensive roadmaps and programs that have accelerated adoption in manufacturing and logistics [6]. SMEs are increasingly using frameworks and models to structure their digital transformation [19], while sectors such as logistics recognize the transformative potential of Industry 4.0 but are still far from fully realizing it in practice [20]. Overall, Industry 4.0 continues to shape the present and near future of manufacturing and logistics, with ongoing efforts to overcome its adoption barriers and to consolidate the benefits of this industrial evolution.

The increasing pace of scientific research and innovation in the fields of Industry 4.0 and manufacturing highlights the urgent need for effective tools to monitor and analyze this progress. This forward momentum is reflected in the steadily decreasing intervals between successive industrial revolutions.

Within this evolving landscape, Industry 5.0 represents a significant advancement in industrial evolution, positioning itself as an essential complement to the concepts developed in Industry 4.0. While the Fourth Industrial Revolution focused on automation, digitalization, and the integration of cyber–physical systems, Industry 5.0 goes beyond by reintroducing the human being as a central component of the production system. This approach aims to establish a synergistic collaboration between humans and machines, addressing critical issues such as environmental sustainability and product customization to meet individual consumer needs [2,21,22,23,24,25,26].

The essence of Industry 5.0 lies in the balanced integration of human capabilities, such as creativity and complex decision-making, with the precision and efficiency of machines. This enables process optimization, fosters innovation, and generates products and services with a highly personalized approach. Technologies such as collaborative robotics, digital twins, machine learning, and the Internet of Things (IoT) are fundamental pillars enabling this new industrial stage [1,25].

A crucial aspect of Industry 5.0 is its focus on three key components: human–machine collaboration, sustainability, and mass personalization. First, human–robot collaboration (cobots) transforms workplace dynamics by allowing workers to focus on creative and supervisory tasks while machines take on repetitive or physically demanding jobs. This approach improves not only efficiency but also working conditions, fostering a safer and more motivating environment [22,24]. Second, sustainability takes center stage through the incorporation of environmentally friendly processes, the adoption of circular economy principles, and the use of green technologies to reduce the carbon footprint [1,26]. Finally, mass personalization, enabled by big data analytics and IoT connectivity, allows products to be tailored to consumers’ specific needs, promoting greater flexibility and customer satisfaction [24,25].

For developing countries, Industry 5.0 offers a transformative opportunity. By combining advanced technologies with a human-centric approach, these nations can overcome structural and economic barriers, foster high-quality jobs, and promote sustainable development. However, this adoption is not without challenges. The main obstacles include digital skills gaps in the workforce, cultural resistance, and limited technological infrastructure [22,25]. These challenges are not exclusive to Industry 5.0 but rather add to or overlap with those inherent in the Fourth Industrial Revolution.

Overcoming these challenges requires a comprehensive strategy that includes educational programs focused on developing technological competencies and public policies that encourage the adoption of advanced technologies. Additionally, fostering alliances between governments, industry, and academia is essential to ensure a successful transition to this new paradigm [1,24].

Against this backdrop, organizations need structured tools to diagnose their current position in the I4.0→I5.0 transition and to plan concrete improvement paths over time. Reference or maturity models and related roadmaps have therefore emerged as key instruments to support both academic analyses and industrial decision-making.

According to Aslanova and Kulichkina [27], Haryanti et al. [28], Teichert [29], digital maturity is a phenomenon that arises with the digital economy and Industry 4.0, understood as the level of adoption and effective use of digital technologies within an organization or sector. However, there is no single, universally accepted definition of this construct, which helps explain the proliferation of heterogeneous maturity models in the literature [27].

Digital maturity not only involves the adoption of advanced technologies but also encompasses organizational development in key dimensions such as processes, culture, and human competencies. According to Li and Lau [30], a comprehensive assessment should include factors such as the alignment of organizational processes and the cultural transformation necessary to adopt new technologies effectively. Similarly, Basl and Doucek [31] emphasize the importance of human resources, leadership, and organizational culture as fundamental pillars for achieving an advanced level of digital maturity. Therefore, a balanced approach integrating these elements is essential for attaining true digital maturity.

Maturity models are strategic and methodological tools designed to assess an organization’s, process’s, or system’s current state of readiness in terms of technological, organizational, and operational capabilities. These models outline a structured progression through stages or levels, reflecting gradual development toward specific transformation and continuous improvement goals [32,33,34]. From a theoretical perspective, maturity models condense and operationalize assumptions about how organizations progress in terms of digital maturity, readiness, and capability development, translating abstract constructs such as digital maturity and Industry 5.0 pillars into observable levels and dimensions [27,28,29,30,31]. From a practical perspective, they are widely used as diagnostic and roadmapping instruments that provide a starting point for planning future development: they help identify gaps, establish strategic goals, and design implementation roadmaps toward greater efficiency, sustainability, and resilience in digital transformation and Industry 4.0/5.0 contexts [32,33,35].

Previous reviews provide important but still partial perspectives on digital and industrial maturity. Teichert [29] synthesizes 22 digital transformation maturity models across sectors, highlighting the fragmentation of maturity dimensions and the emerging role of culture and business models. However, this review does not focus on Industry 4.0 or Industry 5.0 and mainly covers studies published before 2020, that is, before the current proliferation of industrial maturity models. Li and Lau [30] focus on Industry 4.0 maturity models in the domains of information technology and human factors, but they analyse only a small subset of models and primarily emphasize the lack of a standardized model for specific areas such as product safety, rather than providing a comprehensive longitudinal mapping.

Domain-specific reviews also reveal important gaps. Onyeme and Liyanage [33] critically examine 19 smart manufacturing maturity models to assess their applicability to the upstream oil and gas sector. They show that no existing model fully covers the sector’s requirements and that validation is often limited or unclear, which motivates the design of an ad hoc model. Hein-Pensel et al. [34] review Industry 4.0 maturity models for manufacturing SMEs from an Industry 5.0 perspective. Their findings indicate that current models remain predominantly techno-centric and address human-centred aspects only partially, and they call for holistic maturity models that explicitly integrate the three pillars of Industry 5.0. However, their corpus ends in 2022 and does not include models explicitly labelled as Industry 5.0 nor hybrid 4.0/5.0 models.

In light of the above, a specific gap remains: there is no evidence-based longitudinal analysis of how maturity models conceptualize the transition from Industry 4.0 to Industry 5.0 in the most recent period, jointly considering the evolution of model scope (Industry 4.0, hybrid, Industry 5.0), maturity levels, assessed dimensions, and embedded enabling technologies. This study addresses this gap through an Academic Literature Analysis of 75 maturity models published between 2020 and 2024. Specifically, we (i) quantify the temporal evolution from Industry 4.0 models to hybrid and Industry 5.0 models; (ii) propose and apply a reproducible operational definition of “hybrid” models that makes the I4.0→I5.0 transition measurable; (iii) develop a meta-typology of maturity levels and dimensions that explains the persistent dominance of five-level schemes; and (iv) construct an evidence-based cartography of enabling technologies across models. Overall, this review extends previous systematic reviews by providing a longitudinal analysis of the 2020–2024 period, an operational definition of hybrid models at the Industry 4.0–Industry 5.0 interface, and a meta-typology that jointly considers maturity levels, dimensions, and enabling technologies.

From a theoretical perspective, this framework clarifies how the pillars of Industry 5.0, human-centricity, sustainability, and resilience, are layered on top of the digital foundation of Industry 4.0 rather than replacing it. From a practical perspective, the results provide a structured menu of design decisions (level structures, dimension configurations, and technological combinations) that can inform the development of new maturity models, and they help user organizations select or adapt existing models in line with their strategic priorities in terms of efficiency, human factors, and environmental and social performance.

The objective of this article is to compare and analyze maturity models based on levels, dimensions, technologies, and temporal trends in order to (i) advance the theoretical understanding of key differences between existing Industry 4.0 and Industry 5.0 maturity models and (ii) support the practical design and selection of maturity models by user organizations.

Our review is deliberately confined to the maturity models proposed in the peer-reviewed academic literature. In other words, we map the academic design space of Industry 4.0 and 5.0 maturity models as conceptual artefacts, rather than evaluating their diffusion or effectiveness in practice. This choice is consistent with the primary aim of the paper, which is to develop a fundamental metatypology that systematises how models define their scope, dimensions, levels, and underlying assumptions. Focusing on scholarly sources ensures a minimum degree of methodological transparency and conceptual detail, and it provides a traceable corpus where models are documented in sufficient depth to allow systematic comparison. We view this academic metatypology as a necessary precondition for, and complement to, subsequent research that assesses how these models are actually adopted and used in industrial contexts.

The remainder of this paper is structured as follows: Section 2 (Materials and Methods) presents the review design, document retrieval strategy, operationalization of variables, data extraction and coding procedures, and analytical techniques used to address the four research questions. Section 4 (Results) reports the empirical findings on the evolution of Industry 4.0 and 5.0 maturity models, their dimensional coverage, level structures, and embedded enabling technologies. Section 5 (Discussion) interprets these findings in light of the Industry 4.0/5.0 literature and policy agendas, while Section 6 (Limitations and Future Work) summarizes the main constraints of the study and outlines avenues for further research. Section 7 (Conclusions) synthesizes the contributions and practical implications for the design and use of maturity models, and the appendices and open dataset provide the full corpus of 75 models and the GPT-assisted extraction and coding protocol to facilitate replication.

2. Materials and Methods

Four key research questions were established (see

Table 1. Research questions.

Code	Research Question
RQ.1	How have Industry 4.0 and 5.0 maturity models evolved between 2020 and 2024?
RQ.2	Are there significant differences in the evaluated dimensions between Industry 4.0 and 5.0 maturity models?
RQ.3	Is there a significant relationship between the number of levels in Industry 4.0 and 5.0 maturity models and the dimensions evaluated in these models?
RQ.4	What are the most frequently incorporated enabling technologies in Industry 4.0 and 5.0 maturity models?

) to guide the analysis of maturity models. These questions examine (1) the temporal evolution of maturity models within the 2020–2024 timeframe; (2) the main characteristics and differences in maturity dimensions between Industry 4.0 and 5.0 models; (3) the existence, for the purpose of designing new models, of relationships between the number of maturity levels and the number or composition of maturity dimensions; and (4) which enabling technologies support both industrial generations as described in maturity models.

The research was developed through a comprehensive analysis of existing maturity models. The review method is organized into five sequential phases that together form a single end-to-end workflow (see Figure 1). Phase 1—Document retrieval process—identifies and filters the relevant academic sources. Phase 2—Operationalization of variables and categories—defines the constructs, coding rules, and boundary conditions used in the analysis. Phase 3—Data extraction, normalization, and coding—builds the analytical dataset from the selected studies. Phase 4—Analytical procedures—applies the planned analyses to answer RQ1–RQ4. Phase 5—Integration and reporting—synthesizes the findings, links them back to the research questions, and documents the RQ–method–artifact–result traceability summarized in Section 2.5. Each rounded rectangle in Figure 1 represents one of these subprocesses, and the plus symbol indicates that the detailed activities are described in the corresponding subsection (Section 2.1, Section 2.2, Section 2.3, Section 2.4 and Section 2.5).

2.1. Document Retrieval Process (RQ1–RQ4, Review Framework)

The search terms were structured to capture relevant literature using Boolean operators:

("Industry 4.0"’ OR "Fourth Industrial Revolution"’ OR "Industry 5.0"’ OR "Fifth Industrial Revolution"’) AND ("maturity model"’ OR "adoption model"’ OR "‘framework"’)

Web of Science and Scopus were selected due to their recognized coverage of high-quality scientific literature. The method followed for retrieving reference documents is illustrated in Figure 2.

An initial search resulted in 6909 and 7537 publications, respectively. The results were narrowed down by selecting only publications from 2020–2024, applying filters related to relevant topics for the study (see

Table 2. Classifications of Web of Science and Scopus.

Web of Science	Scopus
Engineering Industrial	Engineering
Engineering Manufacturing	Computer Science
Management	Environmental Science
Engineering Electrical Electronic	Energy
Computer Science Information Systems	Economics, Econometrics and Finance
Computer Science Interdisciplinary Applications	Multidisciplinary
Operations Research Management Science
Green Sustainable Science Technology
Computer Science Theory Methods
Environmental Sciences
Computer Science Artificial Intelligence
Business
Telecommunications
Engineering Multidisciplinary
Environmental Studies
Automation Control Systems
Computer Science Software Engineering
Robotics
Economics
Ergonomics
Industrial Relations Labor

), and restricting results to English language and peer-reviewed publications (articles, proceeding papers, and review articles).

Because both databases returned more than 4000 records, we used the built-in “Relevance” ordering in each platform to obtain a manageable but information-rich subset. In Web of Science, we selected the default sorting option “Relevance”, which prioritizes records according to the occurrence and weighting of the query terms in the title, abstract, and author keywords. Similarly, in Scopus, we used the default “Relevance” ordering, which is internally computed from the frequency of the search terms and their weighting by fields (title/abstract/keywords) for the query used. From each database, we exported the top 250 records from this relevance-sorted list (500 records in total before de-duplication). We treat this step as a transparent and reproducible heuristic: any researcher can replicate it by (i) using the same search string, (ii) applying the same temporal, language, and document-type filters (2020–2024; English; articles, conference papers, and review articles), and (iii) sorting by the default relevance criterion in Web of Science and Scopus before exporting the first 250 records from each. After merging both sets, we removed duplicates based on DOI and title, resulting in 444 unique records.

Because the relevance functions of commercial databases are proprietary and may under-represent certain topics (e.g., niche sectors or alternative terminologies), we explicitly addressed potential selection bias in three ways. First, we relied on two independent databases (Web of Science and Scopus), whose relevance algorithms and journal coverage differ, and combined their top-ranked results before screening. Second, we applied broad subject-area filters (engineering, computer science, management, and related fields; see Table 2) rather than very narrow ones, so as not to exclude a priori peripheral but substantively relevant contributions. Third, during full-text screening, we complemented database results with backward reference searching: any maturity model cited as a primary instrument in an included article but missing from the initial set of 444 records was checked manually and, if it met the 2020–2024 inclusion criteria, added to the corpus. We also verified that maturity models for Industry 4.0 considered canonical in prior reviews were retrieved by our search strategy. Taken together, these steps reduce (although they cannot completely eliminate) the risk that relying on relevance algorithms systematically biases the final set of 75 models.

Throughout the screening process, documents were excluded at several stages, as summarized in Figure 2. First, after applying the exclusion criteria (document type, 2020–2024 date range, subject areas, and English language), 2529 records from Web of Science and 2993 from Scopus were discarded. Subsequently, at the relevance-selection stage—in which only the 250 highest-ranked records in each database were retained—an additional 4130 Web of Science documents and 4294 Scopus documents were excluded. Once both sets were combined, 56 duplicate records were removed, leaving 444 unique documents. Full-text screening of these 444 articles led to the exclusion of 206 papers that either did not sufficiently describe Industry 4.0/5.0 maturity models or were not available in full text, resulting in 238 documents for detailed assessment. Finally, during the refinement of the bibliography and data extraction, 163 documents were excluded because they did not provide information that could be operationalized for the analysis, yielding a final corpus of 75 articles. This multi-stage strategy ensured a comprehensive yet operationally feasible corpus focused on high-impact, peer-reviewed literature within the field of interest.

Each of these 75 documents defines a maturity model that is treated as a primary source in the analysis. For full transparency and traceability, the complete list of models (MM.01–MM.75), including their scope (Industry 4.0, hybrid, Industry 5.0), and bibliographic references, is provided in Appendix A (

Table A1. Overview of the 75 maturity models included in the corpus (MM.01–MM.75).

ID	Maturity Model	Scope	Reference	Year
MM.01	RA–RE–RI–RO Maturity Model (Production Management as-a-Service)	Industry 4.0	Abner et al. [53]	2020
MM.02	IMPULS	Industry 4.0	Alcácer et al. [54]	2022
MM.03	Maturity Framework for SMEs in Industry 4.0	Industry 4.0	Amaral and Peças [55]	2021
MM.04	I4.0 MM	Industry 4.0	Angreani et al. [39]	2024
MM.05	Fuzzy Maturity Model	Industry 5.0	Bajic et al. [56]	2023
MM.06	Agca et al. (2017) Maturity Model applied to inventory management	Industry 4.0	Barbalho and Dantas [57]	2021
MM.07	Warehouse 4.0 Maturity Model for SMEs	Industry 4.0	Benmimoun et al. [58]	2024
MM.08	Industry 4.0 with Industry 5.0 aspects (sustainability)	Industry 4.0 & 5.0	Bernhard and Zaeh [48]	2023
MM.09	Maturity Model for SMEs in Industry 4.0	Industry 4.0	Bohorquez and Gil-Herrera [59]	2022
MM.10	ECO Maturity Model	Industry 4.0	Bretz et al. [60]	2022
MM.11	Fuzzy-logic-based Maturity Model for OSCM	Industry 4.0	Caiado et al. [61]	2021
MM.12	Industry 4.0 Maturity Model (Portugal)	Industry 4.0	Castelo-Branco et al. [62]	2022
MM.13	Smart Logistics Maturity Model for SMEs	Industry 4.0	Chaopaisarn and Woschank [63]	2021
MM.14	Maturity Model for Industry 4.0 adoption in Passenger Railway Companies	Industry 4.0	Chaves Franz et al. [64]	2024
MM.15	Modular Maturity Model for Industry 4.0	Industry 4.0	Çinar et al. [65]	2021
MM.16	IPM (Industry 4.0 Perception Maturity)	Industry 4.0	Ciravegna-Martins-da fonseca et al. [66]	2024
MM.17	Maturity Model for General Contractors in Industry 4.0	Industry 4.0	Das et al. [67]	2024
MM.18	S3RM Maturity Model for smart and sustainable supply chains	Industry 4.0 & 5.0	Demir et al. [49]	2023
MM.19	Maturity Model based on TQM	Industry 4.0	Elibal and Özceylan [68]	2024
MM.20	Diagnostics of Opportunities – Maturity Model for Digital Transformation	Industry 4.0	Ericson Öberg et al. [69]	2024
MM.21	Maturity Model for Logistics 4.0	Industry 4.0	Facchini et al. [70]	2020
MM.22	3D-CUBE Readiness Model for Industry 4.0	Industry 4.0	Felippes et al. [36]	2022
MM.23	FITradeoff Industry 4.0 Maturity Model	Industry 4.0	Ferreira et al. [71]	2024
MM.24	Data Science Maturity Model (DSMM)	Industry 4.0	Gökalp et al. [72]	2021
MM.25	Industry X.0 Fuzzy Inference Engine	Industry 4.0	Gomes and Basilio [73]	2024
MM.26	Maturity Model for sPSS	Industry 4.0	Heinz et al. [74]	2022
MM.27	Deloitte-based Digital Maturity Model	Industry 4.0	Herceg et al. [42]	2020
MM.28	Singapore Smart Industry Readiness Index (SIRI) adapted to the Moroccan textile industry	Industry 4.0	Jamouli et al. [75]	2023
MM.29	Upper Austria Industry 4.0 Maturity Model	Industry 4.0	Kieroth et al. [43]	2022
MM.30	Maturity model for digital transformation in the manufacturing industry	Industry 4.0	Kırmızı and Kocaoglu [76]	2022
MM.31	Maturity model to improve company performance through Industry 4.0	Industry 4.0	Koldewey et al. [77]	2022
MM.32	Capability-based maturity model for smart manufacturing	Industry 4.0	Lin et al. [78]	2020
MM.33	Singapore Smart Industry Readiness Index (SIRI)	Industry 4.0	Lin et al. [79]	2020
MM.34	Digital Twin Maturity Model (DTMM)	Industry 4.0	Liu et al. [80]	2024
MM.35	Innovative Capability Maturity Model	Industry 4.0	Lookman et al. [81]	2022
MM.36	COMMA 4.0 (Comprehensive I4.0 Maturity Assessment Model)	Industry 4.0	Lukhmanov et al. [82]	2022
MM.37	Maturity Framework for Readiness toward Industry 5.0	Industry 5.0	Madhavan et al. [83]	2024
MM.38	Maturity model for MSMEs in Industry 4.0	Industry 4.0	Magdalena et al. [84]	2021
MM.39	Industry 4.0 Maturity Index	Industry 4.0	Magnus [85]	2023
MM.40	Lean Smart Maintenance Maturity Model (LSM MM)	Industry 4.0	Maier et al. [86]	2020
MM.41	I4.0 Competency Maturity Model (I4.0CMM)	Industry 4.0	Maisiri et al. [87]	2021
MM.42	Maturity Model for Industry 4.0 integration in any company	Industry 4.0	Melnik et al. [88]	2020
MM.43	Maturity Model for the Autonomy of Manufacturing Systems	Industry 4.0	Mo et al. [89]	2023
MM.44	CUDIE Model (Capability to Utilize Data in Industrial Enterprises)	Industry 4.0	Nausch et al. [90]	2020
MM.45	CCMS 2.0 (Company CoMpaSs) with an AI focus	Industry 4.0	Nick et al. [91]	2022
MM.46	Company Compass (CCMS) 2.0	Industry 4.0	Nick et al. [92]	2021
MM.47	CCMS	Industry 4.0	Nick et al. [93]	2020
MM.48	CCMS2.0e (with AI)	Industry 4.0	Nick et al. [94]	2024
MM.49	TOE-based Digital Maturity Model	Industry 4.0	P. Senna et al. [95]	2023
MM.50	RAISE 4.0	Industry 4.0	Pan Nogueras et al. [96]	2022
MM.51	VPi4 (Industry 4.0 Index)	Industry 4.0	Pech and Vrchota [97]	2020
MM.52	Process Model for the Implementation of Industry 4.0 Use Cases in SMEs	Industry 4.0	Peukert et al. [98]	2020
MM.53	Maturity Model for Machine Tool Companies	Industry 4.0	Rafael et al. [99]	2020
MM.54	Maturity Model for Smart Manufacturing in SMEs in Malaysia	Industry 4.0	Rahamaddulla et al. [100]	2021
MM.55	SSTRA (Smart SME Technology Readiness Assessment)	Industry 4.0	Saad et al. [101]	2021
MM.56	Maturity Model for Urban Smart Factories	Industry 4.0	Sajadieh and Noh [102]	2024
MM.57	LM4I4.0 Maturity Model for manufacturing SMEs in developing countries	Industry 4.0	Sajjad et al. [103]	2024
MM.58	Industry 4.0 Maturity Model Proposal	Industry 4.0	Santos and Martinho [104]	2020
MM.59	Maturity model for Digital Twins in battery cells	Industry 4.0	Schabany et al. [37]	2023
MM.60	Pay-Per-X Maturity Model (PPX)	Industry 4.0	Schroderus et al. [105]	2021
MM.61	Maturity model to assess the impact of Industry 4.0 on SMEs	Industry 4.0	Semeraro et al. [106]	2023
MM.62	I4MMSME Maturity Model for manufacturing SMEs	Industry 4.0	Simetinger and Basl [107]	2022
MM.63	DigiCoM	Industry 4.0	Steinlechner et al. [108]	2021
MM.64	Readiness Assessment of SMEs in Transitional Economies	Industry 4.0	Suleiman et al. [109]	2021
MM.65	Maturity Model for Worker 4.0 adoption	Industry 4.0	Treviño-Elizondo and García-Reyes [110]	2021
MM.66	ECDMM4.0–Employee Competency Development Maturity Model for Industry 4.0	Industry 4.0	Treviño-Elizondo and García-Reyes [111]	2023
MM.67	“Maturity Model to Become a Smart Organization based on Lean and I4.0”	Industry 4.0	Treviño-Elizondo et al. [112]	2023
MM.68	Maturity model for digital twins	Industry 4.0	Uhlenkamp et al. [40]	2022
MM.69	SANOL Industry 4.0 Maturity Model	Industry 4.0	Ünal et al. [113]	2022
MM.70	Industry 4.0 Maturity Model for Manufacturing in India	Industry 4.0	Wagire et al. [114]	2021
MM.71	Maturity model for technological integration in industrial companies	Industry 4.0	Widmer et al. [115]	2022
MM.72	Logistics 4.0 Maturity Model	Industry 4.0	Zoubek and Simon [116]	2021
MM.73	Environmental Maturity Model for Industry 4.0	Industry 4.0	Zoubek et al. [117]	2021
MM.74	Maturity model to assess the automation of production processes	Industry 5.0	Hetmanczyk [118]	2024
MM.75	Maturity model for implementing Logistics 5.0 based on decision-support systems	Industry 5.0	Trstenjak et al. [119]	2022

).

2.2. Operationalization of Variables and Categories

This section concisely and verifiably describes the criteria used to classify and analyze the corpus of 75 maturity models. For each variable, the following elements are specified: (i) operational definition of the construct; (ii) unit of analysis and type of measurement; (iii) classification categories; and (iv) boundary rules with evidence anchors (figures, tables, or instrument excerpts). The objective is to ensure traceability, reproducibility, and comparability among coders, maintaining alignment with the research questions (RQs).

2.2.1. V1—Temporality and Evolution

• Operational definition: Temporal evolution of the model. Enables estimation of relevance, update frequency, and possible I4.0→I5.0 transitions, allowing interpretation of trends and relationships with other variables.
• Unit/measurement: Model; numerical/mixed.
• Recording structure: year_pub (YYYY)
• Boundary rules:
  • If multiple editions exist, the version valid during 2020–2024 is used.
  • In cases with ambiguous dates, the DOI or the repository date of the cited version is prioritized.
• Evidence: Editorial metadata (DOI/date), version notes, forewords, or “Version/Revision History” sections.

2.2.2. V2—Industry 4.0/5.0 Scope

• Operational Definition: This variable captures how each maturity model is positioned along the I4.0→I5.0 transition. It is coded at the model level in one of three mutually exclusive categories: Industry 4.0, Hybrid, or Industry 5.0.
• Unit/Measurement: Model; nominal.
• Coding Principle: The category is assigned based on the evaluative content of the instrument (dimensions, items, indicators, levels), not only on the conceptual labels used in the title or introduction.
• Categories:

  −

  Industry 4.0 (I4.0): Models centred on digital technologies, automation, data and processes/organization (e.g., IoT, CPS, analytics, automation, smart manufacturing), with no explicit indicators on human-centricity, sustainability or resilience. These aspects may be mentioned in the narrative, but they are not part of the scoring structure.

  −

  Hybrid (I4.0 & I5.0): Models that explicitly refer to people, well-being, ethics, sustainability (e.g., SDGs, ESG) or resilience as aims or principles, but that only incorporate them partially or descriptively in the evaluation (e.g., qualitative comments, non-scored criteria). This category also includes models self-labelled as “Industry 5.0” whose instrument remains predominantly techno-centric.

  −

  Industry 5.0 (I5.0): Models that contain at least one dimension, subdimension or group of items that directly evaluates human-centricity (e.g., ergonomics, skills, participation, privacy/security by design), sustainability (environmental and/or social) and/or resilience (e.g., business continuity, flexibility, recovery capacity), and where these aspects are explicitly integrated into levels, indicators or weights.

• Coding Rules:

  −

  The instrument (dimensions, items, level descriptions) is reviewed before assigning the category.

  −

  If explicit indicators on people, sustainability or resilience are present, the model is coded as I5.0, even if the article only refers to Industry 4.0.

  −

  If such indicators are absent, but human-centric, sustainable or resilient goals are stated as principles, the model is coded as Hybrid.

  −

  If neither indicators nor principles related to I5.0 are identified, the model is coded as I4.0.

  −

  When multiple levels of analysis are reported (e.g., plant, line, supply chain), the most granular level for which an evaluable instrument is provided is used as reference.

• Evidence: Coding decisions are anchored in the instrument’s rubrics, evaluation matrices, scales, and criteria tables.

2.2.3. V3—Covered Maturity Dimensions (Classification)

• Operational Definition: Thematic classification evaluated by the instrument. Allows measurement of breadth (covered areas) and balance (technological vs. socio-organizational), and relates to V2.
• Unit/Measurement: Model; multi-label.
• Dimension Catalog:

  −

  Technology/Digital Infrastructure;

  −

  Processes/Operations/Lean CPS;

  −

  People/Competencies;

  −

  Organization/Governance/Leadership;

  −

  Strategy/Business Model;

  −

  Data/Analytics/AI;

  −

  Customer/Value/Experience;

  −

  Sustainability (E/S/G);

  −

  Security/Cybersecurity/Privacy.

• Boundary Rules:
  • A dimension is only coded if it appears in items/criteria/levels.
  • Synonyms are mapped to the corresponding family (e.g., “governance” ↔ organization).
  • In cases of explicit overlap in the instrument, multiple tags are allowed.
• Evidence: Lists of evaluated dimensions, criteria matrices, or level definitions linked to each dimension.

2.2.4. V4—Number of Maturity Levels

• Operational Definition: Formal progression structure. Estimates diagnostic granularity and comparability across models.
• Unit/Measurement: Model; ordinal/categorical (discrete) or continuous (score).
• Categories:

  −

  Discrete Levels: [36] (6, ≥7)

  −

  Continuous/Score: Continuous scale without discrete levels (e.g., 0–1; 0–100)

• Rule: If levels are present but lack distinguishing criteria, code as discrete and add a note on interpretability risk.
• Evidence: Level definitions, evaluation rubrics, thresholds, or progression tables.

2.2.5. V5—Referenced Digital Technologies

• Operational Definition: Technological footprint of the model (within the instrument). Allows profiling of the evaluated stack, detecting coverage biases (e.g., excessive focus on automation vs. data), and linking to V2 and V3.
• Unit/Measurement: Model; multi-label.
• Technology Catalog:

  −

  IoT/CPS: Sensors, IIoT gateways, IT/OT integration, protocols (OPC UA, MQTT), real-time monitoring.

  −

  Cloud/Edge: Cloud/edge deployments, hybrid architecture, service orchestration/provisioning.

  −

  Big Data/Analytics: Data pipelines, quality, lakes/warehouses, descriptive/predictive/prescriptive analytics.

  −

  AI/ML: Applied models (maintenance, quality, planning), MLOps.

  −

  Robotics/AMR/Cobots: Industrial/collaborative robots, AGV/AMR, robotic cells, collaborative safety.

  −

  Simulation/Digital Twin: Process/discrete simulation; connected digital twin (data/state synchronization).

  −

  AR/VR/MR (XR): Operational guidance, training, remote assistance, 3D visualization.

  −

  Additive Manufacturing: 3D printing (polymers/metals), prototyping, direct production.

  −

  Blockchain (DLT): Immutable traceability, smart contracts, data integrity.

  −

  Cybersecurity: Vulnerability management, access control, hardening, continuity, privacy.

• Key Rule: To be marked as “present,” the technology must be integrated into items/criteria/indicators or levels (e.g., checklists, rubrics, KPIs, weighting). Contextual mentions do not count.
• Evidence: Specific items/indicators from the instrument, evaluation tables, or level definitions requiring the technology.

2.3. Data Extraction and Coding (RQ.1–RQ.4)

The process (Figure 3) begins with document collection, incorporating the PDF/article and relevant metadata into a document management system. This is followed by GPT-assisted extraction, using controlled prompts to pre-extract the title, publication year, Industry 4.0/5.0 scope, covered maturity dimensions, and maturity levels, whose output is then subjected to full human verification. Subsequently, the nomenclature of maturity dimensions and enabling technologies is normalized. With the refined matrix, the defined variables are coded, applying boundary rules where applicable.

The instructions for the aforementioned GPT Agent can be found in Appendix B of this article.

2.4. Analytical Procedures

This section aims to document, in a comprehensive and auditable manner, how the data were processed and analyzed to answer each research question (RQ), specifying the analytical techniques used and quality criteria applied. Additionally, it seeks to ensure traceability from Method to Result.

2.4.1. Univariate Descriptive Statistics (RQ.1–RQ.4)

Absolute frequencies and percentages were calculated for categorical variables (e.g., Industry 4.0/5.0 scope, type/number of maturity levels, presence of dimensions and technologies). These measures directly describe the corpus, avoid statistical assumptions, and allow for comparison of category prevalence across models. Summary tables and bar/column charts were used to display category distributions.

• For single-category variables (e.g., scope_i40_i50, levels_type), percentages were calculated based on valid N (i.e., models with available data).
• For multi-label variables (e.g., dim_*, tech_*), a single model may be tagged with multiple categories; therefore, the sum of percentages may exceed 100%.

Each table/figure indicates the total N and, where applicable, the valid n.

2.4.2. Temporal Analysis (RQ.1)

The annual distribution of models (2020–2024) was calculated, along with trend analysis by relevant categories, such as the yearly proportion of Industry 5.0 models or the evolution of dimension and technology coverage. Temporal analysis summarizes, year by year, the presence of approaches and components within the corpus and provides a diachronic reading of their evolution.

• Each model was assigned its year of publication (year_pub).
• In cases of multiple editions, the version valid within 2020–2024 was considered.

Line charts and temporal matrices with counts and percentages per year were employed; each figure specifies the annual denominator (number of models that year) and includes a brief methodological caption.

2.4.3. Descriptive Cross-Tabulations and Exploratory Inferential Analysis (RQ.3–RQ.4)

Cross-tabulation tables (with counts and row/column percentages) were used to explore patterns among variables, for example:

• scope_i40_i50 × dimensions;
• scope_i40_i50 × type/number of levels;
• scope_i40_i50 × technologies.

     These descriptive cross-analyses highlight co-occurrences and profile differences across categories (e.g., whether Industry 5.0 models more frequently report people/ESG dimensions), without performing group-wise mean comparisons. Each table specifies whether percentages are row-based (by scope) or column-based (by dimension or technology). Total N and valid n are reported, and results are shown using tables with totals and percentages, as well as heatmaps that highlight cells with relatively higher or lower presence.

     For RQ.3, we complemented these descriptive summaries with an exploratory inferential analysis to assess the relationship between the number of maturity levels and the breadth of conceptual coverage in terms of dimensions. From the coded matrix, we derived: (i) maturity_levels_num, an ordinal variable capturing the number of explicitly defined maturity levels in each model (3, 4, 5, 6, 10, 11, or continuous scales); and (ii) n_dim_norm, defined as the count of standardized dimension families covered by each model, based on the normalization procedure in Section 2.2. The latter does not represent the raw number of author-declared dimensions but the number of distinct normalized families, and is used as a proxy for the breadth of conceptual coverage.

     Models that did not specify fixed discrete levels or relied solely on continuous scores or qualitative staging were excluded from this inferential analysis. For the Kruskal–Wallis test, models were grouped according to their number of maturity levels. First, we used all available categories (3, 4, 5, 6, 10, 11, 100 levels). Second, to avoid unstable estimates in sparsely populated categories, we estimated a more parsimonious specification restricted to the most common discrete schemes (3–6 levels), excluding outliers with very fine-grained or continuous scales. The null hypothesis ($H_0$) states that the median n_dim_families is the same across groups (no systematic association between number of levels and dimensional breadth); the alternative ($H_1$) states that at least one group differs.

     As a complementary check, we computed the Spearman rank correlation coefficient ($\rho$) between maturity_levels_num and n_dim_families, both for the full set of level-count categories and for the restricted 3–6 level subset. The null hypothesis ($H_0$) posits no monotonic association ($\rho =0$); the alternative ($H_1$) posits a non-zero monotonic association. All inferential analyses were performed in Python (version 3.11.8; Python Software Foundation, Wilmington, DE, USA) using the SciPy library (version 1.11; SciPy Developers/NumFOCUS, Austin, TX, USA). Given the sample size and the design of the review, these tests are interpreted as exploratory diagnostics that complement, rather than replace, the descriptive analysis.

2.4.4. Normalization of Dimensions (V3) and Level Structure (V4)

     In this stage, we analyzed the dimensions assessed by Industry 4.0/5.0 maturity models and the number of levels structuring these models. To ensure comparability across the 75 sources, authors’ terms were grouped and normalized following a structured procedure (see Appendix C).

Dimension Grouping (V3)

• Semantic analysis: evaluation of each dimension’s meaning to detect similarity patterns (keywords and context in the Industry 4.0 ecosystem). For instance, “Cybersecurity” and “Security and Governance” were grouped due to their shared aim of protecting systems, data, and processes.
• Thematic classification: identification of broad categories (technology, processes, people, sustainability, etc.) as transversal axes. Terms such as “Technology”, “Digitalization”, and “IoT” were grouped under Technology and Digitalization.
• Functional relationships: assessment of how each dimension contributes to broader goals; e.g., “Smart Processes” and “Lean Production” grouped under Processes and Operations (operational efficiency).
• Hierarchical abstraction: association of specific dimensions with broader categories according to specificity; e.g., “Workforce Training” and “Human Resource Development” within People and Competencies.
• Alignment with reference frameworks: consideration of recognized pillars in the literature and frameworks (e.g., RAMI 4.0) to structure categories into strategic areas; e.g., “Big Data” and “Artificial Intelligence” in Technology and Digitalization.
• Intersection and multidimensionality: when a term could fit multiple categories, assignment favored the primary purpose, acknowledging cross-cutting concepts such as Sustainability and Social Responsibility.
• Parsimony and clarity: unification of near-synonyms to avoid redundancy; e.g., “Material Flow Automation” and “Automation” normalized under Technology and Digitalization.

Level Structure Coding (V4)

We recorded for each model whether levels were discrete or continuous and the number of levels reported. These codifications were used in descriptive summaries and cross-tabulations with model scope (I4.0/Hybrid/I5.0) and publication year (2020–2024) for later analysis.

Outputs for Subsequent Analysis

The standardized taxonomy and the coded variables (V3, V4) served as inputs for frequency-by-year trends, contingency summaries (scope × dimensions and scope × levels), and visualizations described in Section 4.

2.4.5. Visual Representations (RQ1–RQ4)

The following types of tables and charts were used to represent each result or resource supporting the analysis:

• Bar Charts (simple, stacked, or grouped): Used to show differences in prevalence between categories.

  −

  Stacked bars describe the internal composition of each category.

  −

  Grouped bars present side-by-side comparisons across subcategories.

  Charts are typically sorted from highest to lowest frequency. Labels are expressed as percentages when the denominator corresponds to each group’s total, or as raw counts when absolute volume is of interest.
• Line Charts (2020–2024 series):

  −

  These are used to depict year-over-year changes with a single axis.

  −

  The same axis range is maintained across comparable charts.

  −

  The number of models per year (annual N) is indicated.

  Dual axes and 3D effects are not used.
• Matrices/Heatmaps:

  −

  These summarize patterns in descriptive cross-tabulations with multiple category combinations.

  −

  The color scale is adjusted by row or column based on the reported percentage and includes a legend.

All figures include axes with units (counts/percentages) and self-explanatory titles. The sample size Nand valid n (when applicable) are explicitly noted. Numbering follows a consistent scheme (Figure 1..N; Table 1..N) with uniform styles. Figure captions include a brief methodological note (e.g., “counts and percentages; 2020–2024 annual series”), the variable(s) used and, where necessary, a reference anchor to the coding matrix.

2.5. Integration and Reporting

This subsection integrates the four research questions (RQ1–RQ4) with the analytical variables, methods, and outputs, thereby providing the core conceptual framework of the study. Rather than adding a separate model detached from the empirical workflow, we formalize the design as a traceability artefact that links: each research question; the variables through which it is operationalized (scope, dimensions, levels, and technologies); the analytical procedures applied; and the concrete outputs reported in Section 4.

Table 3. MR map (research question–variable–method–artifact–result).

RQ	Variable	Procedure (Methods)	Evidence/Artifact	Linked Result
RQ.1 Evolution of scope (I4.0/Hybrid/I5.0)	V1; V2; V3	Univariate Descriptive Statistics Temporal Analysis Normalization of dimensions and level structure Visual Representations	Figure 4 Figure 5 Table 4 Table 5 Figure 6	Temporal transition pattern from I4.0 to I5.0 and category distribution
RQ.2 Dimension coverage	V2; V3	Univariate Descriptive Statistics Normalization of dimensions and level structure Visual Representations	Table 6	Coverage breadth and dimensional profiles
RQ.3 Level structure	V2; V3; V4	Univariate Descriptive Statistics Normalization of dimensions and level structure Descriptive Cross-Tabulations Visual Representations	Table 7 Table 8	Distribution of discrete/ continuous levels and comparisons by scope
RQ.4 Referenced technologies	V2; V5	Univariate Descriptive Statistics Descriptive Cross-Tabulations Visual Representations	Figure 7	Technological profiles by scope category

functions as this integration map. For each research question, it specifies the subset of variables used: RQ1 combines temporality and scope (V1–V2) with the normalized dimension taxonomy (V3) to characterise the evolution of Industry 4.0, hybrid, and Industry 5.0 maturity models over time. RQ2 focuses on the conceptual content of models by crossing scope (V2) with maturity dimensions (V3). RQ3 extends this by incorporating the level structure (V4), thus relating scope, dimensions, and number/type of maturity levels. RQ4 links scope (V2) to the referenced enabling technologies (V5), yielding technological profiles for each industrial generation. In this way, the table makes explicit how the four core design variables of the study, scope, dimensions, levels, and technologies, are systematically mobilized to answer each RQ.

The third column of Table 3 summarises the methodological building blocks applied to each RQ (descriptive statistics, temporal analysis, cross-tabulations, exploratory inferential tests, and visual representations), while the fourth and fifth columns bind these procedures to specific artefacts (tables and figures) and to the type of result they support. Taken together, these links ensure end-to-end traceability from research questions to coded variables, methods, and reported outputs. Conceptually, the table therefore operates as an integration framework that connects the review questions with the scope of the models (Industry 4.0, hybrid, Industry 5.0), their internal design features (maturity dimensions and level structures), and their technological footprint, rather than as an isolated checklist. This explicit RQ–variable–output mapping clarifies how the different components of the study cohere into a single analytical design.

3. Declaration of Generative AI and AI-Assisted Technologies in the Writing Process

During the preparation of this work, the authors utilized the artificial intelligence ChatGPT (OpenAI GPT-4o, 2024 release) for reasoning models (O1, O3), data processing, and the improvement and correction of scientific writing in English (4o), as well as the GPT assistant MMs4.0, designed by the authors to extract and structure data. After using these tools, the authors reviewed and edited the content as necessary and assume full responsibility for the publication’s content.

4. Results

4.1. RQ.1 How Have Industry 4.0 and 5.0 Maturity Models Evolved Between 2020 and 2024?

From the total number of analyzed models (see Figure 4), 92% (69 models) correspond to Industry 4.0 maturity models, while 5% (4 models) are associated with Industry 5.0. Additionally, 3% of the models refer to both Industry 4.0 and Industry 5.0.

This result reflects the predominant focus on Industry 4.0, given its longer history and the broad recognition of its impact on companies’ digital transformation. In contrast, Industry 5.0, being more recent, is still in an emerging phase but has shown growth in publications in recent years.

In addition to the previous analysis, the temporal behavior of model volume (Figure 5) reveals that the number of published maturity models … related to Industry 4.0 was higher between 2020 and 2024, peaking in 2021, when 18 Industry 4.0 models were published. The year 2022 also saw a significant number of publications, with 17 Industry 4.0 models.

However, in 2023, a decline in the number of published Industry 4.0 models was observed, with only 8 models recorded. Nevertheless, 2023 was also the year when the first combined Industry 4.0 and 5.0 models emerged (2 models), along with the first exclusive Industry 5.0 model.

In 2024, there was a slight increase in the publication of Industry 5.0 models, with a total of 2 additional models, while the number of Industry 4.0 models remained at 13. This suggests a growing interest in Industry 5.0 in the past year.

Maturity Dimensions

As observed in

Table 4. Frequency of maturity model dimensions.

Dimension	Frequency	%
Quality	4	5%
Customers and Market	7	9%
Management and Strategy	38	51%
Infrastructure and Ecosystem	16	21%
Innovation and Value Creation	27	36%
Human–Machine Interaction	4	5%
People and Competencies	34	45%
Processes and Operations	38	51%
Security and Governance	9	12%
Sustainability and Social Responsibility	10	13%
Technology and Digitalization	53	71%

, Technology and Digitalization is the most frequent dimension, present in 71% of the models, followed by Management and Strategy (51%) and Processes and Operations (51%).

On the other hand, Sustainability and Social Responsibility, although currently a relevant dimension, appears in only 13% of the models, suggesting that it is not yet widely considered in many maturity frameworks, similar to Customers and Market, Human–Machine Interaction, and Quality, with 9% and 5%, respectively.

Regarding the number of dimensions structuring the models, they are typically divided into categories that allow evaluation of a company’s progress in adopting Industry 4.0 and 5.0 technologies and practices. In this sample, models are distributed across eight different dimension configurations: 2, 3, 4, 5, 6, 7, 8, and 12 dimensions (see

Table 5. Frequency of number of dimensions per maturity model.

Number of Dimensions	Frequency	% of Total
2	2	2.67%
3	17	22.67%
4	20	26.67%
5	16	21.33%
6	12	16.00%
7	5	6.67%
8	2	2.67%
12	1	1.33%

). Most maturity models are concentrated in those that include 3, 4, and 5 dimensions, which together represent 71% of the total, suggesting that this level of granularity is generally seen as sufficient to capture a comprehensive view of industrial digital maturity without overcomplicating the analysis.

Models with more than seven dimensions are rare, indicating that few frameworks attempt to encompass an extremely detailed analysis. The most extensive model evaluates 12 dimensions [37] in a highly specific sector and technological context (digital twins in battery cell manufacturing). In that case, the larger number of dimensions is justified by the need to capture both technical and organizational aspects in a complex environment. The authors report that the decision to include 12 dimensions was refined through workshops and peer comparisons to obtain a holistic yet still manageable framework, enabling a thorough evaluation of maturity aspects without introducing unnecessary complexity.

General Trends by Dimension (2020–2024)

The temporal behavior of the standardized dimensions (Figure 6) shows that some of them are consistently central in maturity models throughout the 2020–2024 period, while others emerge more recently or remain marginal. Technology and Digitalization is the dominant dimension every year and reaches its highest frequency in 2021–2022, which is consistent with the fact that both Industry 4.0 and Industry 5.0 build on a strong digital and data-driven foundation. Management and Strategy and Processes and Operations also appear very frequently in 2020–2022, reflecting an early focus on strategic alignment and process optimization as core levers of digital transformation. Both dimensions decrease noticeably in 2023 and partially recover in 2024, but without returning to their previous peak.

Other dimensions exhibit a growth pattern that is particularly visible in the first half of the period. People and Competencies, which captures human talent and skills, increases between 2020 and 2022 and becomes one of the most cited dimensions in 2022, indicating that more recent models increasingly recognize the role of the workforce and training as enablers of digital transformation. A similar trajectory is observed for Innovation and Value Creation, which also peaks in 2022, suggesting that as digital initiatives mature, greater emphasis is placed on their contribution to business value and innovation rather than on basic adoption alone. Both dimensions decline in 2023, in line with the lower overall number of models published that year, but remain clearly present in 2024.

Two dimensions are practically absent at the beginning of the period but later become visible: Sustainability and Social Responsibility and Human–Machine Interaction. Sustainability is not mentioned in any of the 2020 models but appears from 2021 onwards and remains present in subsequent years, signalling that environmental and social concerns have begun to be explicitly incorporated into maturity assessments. Human–Machine Interaction, which refers to the collaboration between workers and robotic or smart systems, also appears for the first time in 2021 and then reappears in later years. The timing of these emergences coincides with the diffusion of the Industry 5.0 paradigm (around 2021–2022), which emphasizes human-centric and sustainable approaches on top of the Industry 4.0 digital backbone.

By contrast, dimensions such as Security and Governance, Customers and Market and Quality maintain low frequencies throughout the period and do not exhibit a clear increasing or decreasing trend. This may indicate that, in many models, security, customer focus and quality are addressed implicitly within other dimensions (for example, processes, technology or strategy) rather than as standalone pillars, or that they are considered less distinctive for differentiating maturity levels in the context of Industry 4.0/5.0.

Figure 6 summarizes these patterns by showing, for each year, how many models include each standardized dimension. The vertical axis represents the number of models and the horizontal axis the year of publication. Technology and Digitalization remains the highest line across the series, Management and Strategy and Processes and Operations follow a similar rise-and-fall behavior around 2023, People and Competencies and Innovation and Value Creation show a build-up towards 2022, and Sustainability and Social Responsibility and Human–Machine Interaction start at zero in 2020 and then appear modestly from 2021 onwards.

From a year-by-year perspective, 2020 largely reflects the “classical” priorities of Industry 4.0: most models published in that year include Technology and Digitalization, Management and Strategy and Processes and Operations, while dimensions related to sustainability, human–machine interaction or broader societal concerns are entirely absent. In 2021 and 2022, models become more multidimensional: the total number of dimension mentions reaches its maximum, Sustainability and Human–Machine Interaction appear for the first time, and People and Competencies and Innovation and Value Creation strengthen markedly. In 2022, in particular, technological, organizational, human and emergent 5.0-related aspects converge, and the first model explicitly labelled as Industry 5.0 is documented, reinforcing the interpretation of this year as a peak of conceptual integration.

In 2023, there is a clear inflection: the number of models and the total count of dimension mentions drop, and most dimensions decrease in frequency. This contraction coincides with the emergence of hybrid Industry 4.0/5.0 models and one model explicitly framed as Industry 5.0, which tend to give more weight to sustainability while not always retaining all the traditional Industry 4.0 dimensions (such as Quality or Customers and Market). In 2024, the volume of models increases again and core digital dimensions regain prominence, while human- and sustainability-related dimensions remain present but are not yet universal. Overall, the temporal pattern suggests an evolution from a technology- and process-centric view of maturity (2020) towards more complex and multidimensional frameworks (2021–2022), followed by a phase of consolidation and rebalancing (2023–2024) in which Industry 4.0 remains numerically dominant but Industry 5.0 pillars begin to be integrated more explicitly.

The vertical axis shows the number of models that include each dimension, while the horizontal axis corresponds to the year of publication.

As shown in Figure 6, Technology and Digitalization (black line) is the most prevalent dimension across all years, maintaining a high and relatively stable presence. Management and Strategy (orange line) and Processes and Operations (gray line) also dominate in 2020–2022 but decline in 2023. People and Competencies (yellow line) and Innovation and Value Creation (green line) show significant increases until 2022, followed by declines. On the other hand, Sustainability and Social Responsibility (pink line) and Human–Machine Interaction (overlapping with other lines) start at zero in 2020 and then rise modestly from 2021 onward, illustrating their late emergence. Dimensions such as Security and Governance (purple line), Customers and Market (brown line), and Quality (red line) remain low and flat, with minor fluctuations or missing years.

The following is a comparison of the dimensions highlighted in each year, and notable changes or events are pointed out: The year 2020 is the first analyzed and largely reflects the classic priorities of Industry 4.0. The models published in 2020 primarily emphasize Technology and Digitalization, Management/Strategy, and Processes/Operations (each present in 9 models). In other words, almost all models include these three core dimensions of digital transformation. Other dimensions, such as Innovation and People, appear but less frequently (4 times each), suggesting that some early models did not explicitly consider them. Social or environmental dimensions (Sustainability, Human–Machine Interaction) were completely absent in 2020.

In 2021, there is a general increase in the complexity and scope of the dimensions covered. This year introduces Sustainability and Social Responsibility (3 models) and Human–Machine Interaction (2 models) for the first time, marking the beginning of new concerns beyond the “Fourth Industrial Revolution.” Additionally, almost all dimensions increase their frequency compared to 2020: for example, People and Competencies rises to 8 mentions (doubling its presence), Innovation increases to 6, Infrastructure/Ecosystem rises to 4, while Technology, Strategy, and Operations remain very high (10 mentions each, even slightly increasing). Clearly, 2021 was a year when maturity models incorporated more dimensions per model, signaling a more holistic approach. This change can be attributed to the evolution of Industry 4.0 understanding and the early emergence of Industry 5.0 in early 2021. Notably, the introduction of Sustainability as a dimension in 2021 coincides with the publication of the European Union’s Industry 5.0 report [38], which emphasized integrating sustainable and human-centered goals into the industry.

The year 2022 has the highest total number of mentioned dimensions (64 occurrences in total, see table), suggesting that many models were published and/or each model included numerous dimensions. Several dimensions peak in 2022: Technology and Digitalization reaches 13 mentions (present in almost all models that year), People and Competencies peaks at 11, Innovation and Management and Strategy also reach their highest frequency (10 each). This indicates that by 2022, maturity models strongly integrated technological, organizational, and human capital aspects. Processes and Operations curiously decreases slightly to 8 (compared to 10 the previous year), possibly reflecting that some newer models prioritized innovation or people over traditional processes. Sustainability and Social Responsibility consolidates at 3 mentions (similar to 2021, suggesting it remained an additional component in some models but was not universally included). Human–Machine Interaction appears in 1 model in 2022. This year also marks the first documented case of a model explicitly classified as “Industry 5.0,” confirming the influence of this emerging paradigm. Overall, 2022 represents the peak of multidimensional integration: maturity frameworks encompass technology, sustainability, strategy, people, and innovation.

In 2023, a notable shift is observed: the number of models or dimensions sharply declines (only 33 total occurrences for the year). Consequently, many dimensions show significant decreases in frequency. For example, Management and Strategy drops from 10 to merely 3 occurrences (a substantial reduction), Innovation from 10 to 3, People from 11 to 5, Technology from 13 to 8, and Processes from 8 to 6. Practically all dimensions decrease in 2023, except for Security and Governance (remaining at 2) and Sustainability and CSR (holding steady at 3, the same as the previous two years). Some dimensions even temporarily disappear: Quality and Customers/Market do not appear in any models that year. What could have happened in 2023? One possible explanation is the fewer publications of new models (only 8 purely Industry 4.0 models identified that year, compared to 17–18 in previous years), perhaps due to maturity model research reaching saturation between 2020–2022 or focusing on refining existing models instead of creating new ones. Another explanation is the emergence in 2023 of 2 models classified as “Industry 4.0 and 5.0” and 1 exclusively “Industry 5.0” model, which tend to include dimensions such as Sustainability but possibly omit some traditional 4.0 dimensions (for instance, these models might exclude Quality or Customers in favor of 5.0 dimensions). Indeed, the fact that Sustainability maintains 3 mentions in 2023 (not decreasing) suggests almost all models published that year incorporated sustainability, consistent with several models already embracing the 5.0 approach. In summary, 2023 marks an inflection point with fewer models more oriented toward Industry 5.0, resulting in decreased emphasis on strategy, innovation, and people (in terms of the number of models including them) and continued focus on technology and sustainability in the few available models.

The data from 2024 (up to the research compilation) show some recovery in the number of models (40 dimension occurrences, up from 33 in 2023). Certain dimensions slightly rebound: Management and Strategy rises to 6, People to 6, and Technology again to 10. This might indicate that maturity model research continued, potentially blending elements of Industry 4.0 and 5.0. However, Sustainability falls to 1 mention in 2024 (from 3 in previous years). This decline could be because the few identified 5.0 models were already counted in 2023, and in 2024, most collected models reverted primarily to Industry 4.0 (13 Industry 4.0 models versus only 1 Industry 5.0 model in 2024, according to data). Thus, while 2024 recovers in volume, the emphasis on sustainability was not as prominent as in 2021–2023, perhaps indicating that not all new models systematically incorporate it yet. Other minor dimensions remain low (Quality at 1, Customers at 1, Human–Machine Interaction at 1). The year 2024 suggests a convergence: models continue to focus on digital aspects (technology, processes) while moderately integrating human factors (people, strategy), and some include sustainability, although not universally.

4.2. RQ.2 Are There Significant Differences in the Dimensions Evaluated Between Industry 4.0 and 5.0 Maturity Models?

Correlation with Industry 4.0 vs. Industry 5.0. The data provide (see

Table 6. Frequency of evaluated dimensions in Industry 4.0 and Industry 5.0 maturity models.

Dimensions	Industry 4.0	Industry 4.0 & 5.0	Industry 5.0	Total
Quality	4	-	-	4
Customers and Market	7	-	-	7
Management and Strategy	38	-	-	38
Infrastructure and Ecosystem	15	-	1	16
Innovation and Value Creation	26	-	1	27
Human–Machine Interaction	4	-	-	4
People and Competencies	33	-	1	34
Processes and Operations	35	1	2	38
Security and Governance	9	-	-	9
Sustainability and Social Responsibility	8	2	-	10
Technology and Digitalization	50	1	2	53

) insights into how the transition from Industry 4.0 to Industry 5.0 has influenced the dimensions of maturity models:

• Technological vs. human-sustainable focus: Models classified as Industry 4.0 (predominant until 2021–2022) focused heavily on digitization, automation, and efficiency. This explains why dimensions such as Technology, Processes, and Strategy dominated those years. Industry 4.0 involved extensive digitization and automation of production using IoT, AI, robotics, etc., aiming for increased productivity and real-time decision-making. In contrast, Industry 5.0 (conceptualized from 2021 onwards) complements 4.0 by emphasizing environmental sustainability, resilience, and human-centricity. Consequently, dimensions directly related to these new emphases (Sustainability, Human–Machine Interaction) only appear as the Industry 5.0 agenda takes hold. This reinforces the correlation: it is not coincidental that these dimensions emerged prominently from 2021–2022; they reflect the new vision encouraging industry to adopt technology in a human-centric and sustainable way. For example, Industry 5.0 promotes “optimizing Human–Machine Interactions” and “making production respectful of planetary boundaries” [38], precisely what the Human–Machine Interaction and Sustainability dimensions evaluate in a maturity model.
• Progressive integration: Initially, models were purely Industry 4.0 (all models in 2020 and most from 2021–2022). However, even before explicit “5.0” models existed, some models from 2021 already included sustainability within a 4.0 context. This suggests a gradual transition: authors of 4.0 models voluntarily began incorporating sustainability and human factors, anticipating their growing relevance. The first explicitly Industry 5.0 model emerged in 2022, and by 2023, more hybrid 4.0/5.0 or pure 5.0 models appeared, solidifying these dimensions. The increasing prominence of People and Competencies also aligns with Industry 5.0’s vision of empowering workers rather than replacing them, highlighting employee training and well-being as keys to maturity [38]. This indicates that the boundary between 4.0 and 5.0 in maturity models is fluid: during 2021–2022, elements of both paradigms coexisted until formal 5.0 models emerged.
• Persistence of 4.0 foundations: Despite the arrival of new approaches, core Industry 4.0 dimensions remain intact. Technology and Digitalization continue to be omnipresent (even Industry 5.0 models include them since technological foundations remain essential). Processes and Operations, and Management and Strategy, though fluctuating, still feature prominently in many models. This aligns with the notion that Industry 5.0 complements rather than replaces 4.0. That is, Industry 5.0 advancements add layers (sustainability, resilience, human-centric approach) built upon the platform established by 4.0 (digitization, automation). For example, an Industry 5.0 maturity model evaluates both how digitized the company is (a legacy dimension from 4.0) and how sustainable or human-centered it is (new dimensions from 5.0). In summary, coexistence is evident: during 2023–2024, the few Industry 5.0 models still assess technological and process maturity while incorporating new criteria.
• Shifts in priority: However, adjustments in relative importance are noticeable. In recent years, driven by Industry 5.0, sustainability transitioned from nonexistence to prominence in advanced models, while dimensions like Quality lost explicit relevance. This suggests models adapt their focus according to current priorities. Quality was foundational in traditional industrial models but might now be implicitly assumed or integrated within operations in the digital-sustainable era. Conversely, sustainability emerges as a new cornerstone. Thus, the evolution of dimensions reflects the industry’s shifting goals: from seeking efficiency and quality (Industry 4.0) toward additionally emphasizing sustainability and resilience (Industry 5.0). As a reference, the European Commission highlights that industry should now “accelerate innovative change” and provide prosperity “while respecting planetary boundaries and placing worker well-being at the center,” a message clearly influencing maturity models post-2021 [38].

When examining differences between Industry 4.0 and 5.0 models more deeply, significant variations emerge in the evaluated dimensions (see Table 6). Industry 5.0 models tend to emphasize human-centric dimensions such as Sustainability and Social Responsibility and Human–Machine Interaction. This reflects Industry 5.0’s focus on greater integration between humans and intelligent systems and growing concerns about sustainable and ethical practices.

In contrast, Industry 4.0 models predominantly emphasize technical and organizational dimensions such as Technology and Digitalization, Processes and Operations, and Management and Strategy. Although these dimensions are also present in Industry 5.0 models, their priority and emphasis differ, highlighting an evolution in objectives and approaches between these two industrial revolutions.

4.3. RQ.3 Is There a Significant Relationship Between the Number of Levels in Industry 4.0 and 5.0 Maturity Models and the Dimensions Evaluated Within These Models?

Analyzing maturity models concerning the number of maturity levels reveals trends and patterns in their structural design (see

Table 7. Frequency of the number of maturity levels.

Maturity Levels	Frequency	% of 75 Models
3 levels	6	8.00%
4 levels	8	10.67%
5 levels	37	49.33%
6 levels	14	18.67%
10 levels	1	1.33%
11 levels	1	1.33%
100 levels (percentage scale)	1	1.33%
No fixed levels/alternate method	7	9.33%

). Of the 75 models evaluated, 68 explicitly specify or define the number of maturity levels, representing approximately 90.7% of the total. Conversely, 7 models, or 9.3%, do not detail or define their number of levels, potentially affecting their interpretation and practical application.

Among the 75 models, 6 (8.0%) present three levels, 8 (10.7%) have four levels, and the majority, 37 (49.3%), adopt a five-level structure. Additionally, 14 models (18.7%) possess six levels, while 1 model (1.3%) has ten levels, 1 model (1.3%) has eleven levels, and another model (1.3%) employs a 0–100 percent scale. Finally, 7 models (9.3%) do not define fixed discrete levels and instead rely on alternative approaches such as continuous scores or qualitative staging; these designs are comparatively atypical within the sample.

Between 2020 and 2024, most maturity models maintained similar structural patterns, predominantly featuring five-level models throughout the period. The consistent preference for a five-level structure suggests a stable inclination toward this standard format. There was no drastic annual shift observed in terms of the number of levels used: models published in 2020 already favored five levels, and this practice has continued through 2024.

The prevalence of five-level models indicates a common trend in maturity model design. It can be inferred that this structure provides an appropriate balance between granularity and simplicity necessary for assessing organizational maturity progress. This configuration allows for clear and practical differentiation between developmental stages without overly complicating the model.

Variability in the number of levels among different models reflects adaptation to specific needs and contexts. For instance, the model presented by [39] illustrates that the number of maturity levels in Industry 4.0 models correlates directly with the depth and complexity of the digital transformation being evaluated. Their model shows that a three- or four-level structure provides organizations with a solid foundational assessment, from basic data collection to initial integration, enabling a general evaluation without requiring exhaustive process analysis. This approach is particularly suitable for companies in early transformation stages or those seeking an overarching view without delving into overly specific details.

Conversely, the model also demonstrates that when more precise evaluations are required, structures with six or more levels are preferred. A higher number of levels allows for finer differentiation regarding technology integration, advanced automation, and intelligent service provision, crucial for organizations advanced in their digital maturity. This detailed approach identifies specific improvement areas and outlines precise progression paths, adapting to complex contexts where each step in the transformation process demands rigorous monitoring.

The absence of detailed level information in a notable proportion of models mainly occurs because many approaches prioritize qualitative and descriptive evaluations of capabilities and practices rather than numerical or discrete scales. For instance, the Digital Twins model by [40] assesses maturity using continuous scores, calculated via membership functions and weighting, instead of fixed levels, providing flexibility to accommodate diverse use cases. Similarly, the model by [41] structures its solution hierarchically (strategic, tactical, and operational planning), responding to specific organizational roles and responsibilities rather than proposing a progressive maturity scale. Furthermore, some models, such as the one discussed by [42], emphasize qualitative aspects, describing capabilities or practices without assigning specific levels, thereby supporting the absence of fixed scales and enabling adaptation to organizational specifics.

The exceptional case of model MM.29 [43], featuring eleven levels, offers highly detailed maturity classification, enabling precise differentiation among evaluated states. Although this highly granular approach increases evaluation complexity, it proves especially beneficial in contexts of high uncertainty and heterogeneity, where precise identification of improvement areas and incremental advancement paths is crucial. Similarly, employing ten or even a hundred levels (0–100 scale) in other models highlights equally uncommon approaches oriented toward fine or continuous measurement.

These observations have significant implications for maturity assessment. The dominance of five-level models might facilitate comparability across evaluations and promote industry standardization. However, existing variability indicates that maturity models remain adaptable and customizable according to sector-specific or organizational needs.

The lack of detailed specification in many models underscores the importance of clearly defining the maturity model structure. A precise definition of levels is essential for ensuring model usability and applicability, facilitating the understanding and interpretation of results. Therefore, it is advisable for maturity model developers to explicitly specify the number and description of levels.

Organizations should consider which level structure best suits their needs, balancing simplicity with depth in evaluation. When using or comparing maturity models, it is crucial to account for the number of levels and how this influences result interpretation. A clear understanding of model structure enables organizations to make more informed decisions regarding implementation and monitoring.

4.3.1. Relationship Analysis Between Dimensions and Proposed Maturity Levels

The matrix represented in

Table 8. Relationship matrix between dimensions and number of maturity levels.

Dimensions/Levels	0	3	4	5	6	10	11	100	Total
Quality	1	1	-	2	-	-	-	4	8
Customers and Market	-	2	1	3	1	-	-	-	7
Management and Strategy	-	-	-	20	11	-	-	-	33
Infrastructure and Ecosystem	1	2	1	8	3	-	1	-	16
Innovation and Value Creation	-	3	1	13	8	-	-	-	27
Human–Machine Interaction	-	-	1	2	1	-	-	-	4
People and Competencies	-	3	2	20	6	-	1	-	32
Processes and Operations	-	2	5	20	9	-	-	-	38
Security and Governance	-	-	-	5	4	-	-	-	9
Sustainability and Social Responsibility	-	-	-	8	2	-	-	-	10
Technology and Digitalization	-	-	3	22	12	1	-	-	38

establishes a relationship between the dimensions identified in the evaluated maturity models and the number of maturity levels present in these models. The following analysis focuses on the relationships between the variables “Number of Levels” and “Dimensions,” identifying significant patterns and trends emerging from the data.

As previously confirmed, the dimension Technology and Digitalization is the most recurrent, with a total of 53 occurrences in the analyzed models. Its presence is particularly notable in models with 5 levels (24 times) and in models with 6 levels (12 times). This is not surprising, as it indicates that regardless of the complexity or specificity of the model, technology and digitalization are considered fundamental pillars in assessing Industry 4.0 and 5.0 maturity. Its high frequency corroborates that these areas are critical for both current and future industrial competitiveness.

Similarly, Processes and Operations has a strong presence in 5-level models (20 times) and is also significant in 6-level models (9 times). This reflects that internal process optimization and efficiency are key aspects in most maturity models, making their assessment essential for measuring organizational performance. Also appearing 38 times, Management and Strategy stands out among 5- and 6-level models, occurring 20 and 11 times, respectively. The next most important dimension is People and Competencies, again most commonly reflected in 5- and 6-level models, with 20 and 6 occurrences, respectively. This highlights the importance of clear strategic direction and human capital development in an organization’s maturity. Effective management and employee competencies are critical factors directly influencing an organization’s ability to adapt and grow in dynamic environments.

The dimension Innovation and Value Creation appears 27 times, most frequently in 5-level models (13 times). This suggests that innovation is seen as a key driver of value creation and that its evaluation is more detailed in models with a greater number of levels, allowing for a more granular appreciation of how organizations innovate and generate competitive advantages.

When examining the relationship between the number of levels and dimensions, models with 5 levels tend to include a greater variety and frequency of dimensions. This could indicate that a 5-level structure provides an optimal balance between detail and manageability, enabling a comprehensive evaluation without excessive complexity. Core dimensions such as technology, processes, management, and people are consistently highlighted in these models, reinforcing their central relevance.

In contrast, models with 3 and 4 levels show lower frequency in dimension occurrences. For example, Technology and Digitalization appears 7 times in 3-level models and 5 times in 4-level models. This could suggest that models with fewer levels group dimensions into broader categories or focus on more general aspects of maturity, sacrificing detail for simplicity.

Models with 6 levels exhibit an intermediate distribution. The Technology and Digitalization dimension appears 12 times, and other dimensions such as Management and Strategy and Processes and Operations also have a significant presence. This suggests that these models aim for a higher level of detail than those with fewer levels, possibly breaking down specific aspects within key dimensions.

Notably, the Sustainability and Social Responsibility dimension appears 10 times, being most frequent in 5-level models. This reflects a growing awareness and appreciation of sustainability in organizational maturity assessments. The inclusion of this dimension suggests that models guide organizations toward recognizing the importance of sustainable and ethical practices, not only for regulatory compliance but also as a strategic component for long-term success.

The Human–Machine Interaction dimension has a lower overall frequency (4 times), mainly appearing in 5-level models. Although less represented, its presence indicates an emerging focus on effective integration between humans and automated technologies, a relevant aspect in the context of Industry 4.0 and 5.0.

The Quality dimension, with only 4 occurrences, shows limited presence. This could be interpreted in various ways: quality may be implicitly included in other dimensions such as processes and technology, or some models may prioritize other aspects in maturity evaluation. However, its low frequency highlights a potential opportunity to reinforce the importance of quality in future models.

For models with unspecified levels, key dimensions still maintain a significant presence. This suggests that while the level structure is not traditionally defined, the core dimensions remain relevant for maturity assessment. However, the lack of specification could limit evaluation precision and comparability.

The particular case of the model with 11 levels shows occurrences in dimensions such as Technology and Digitalization and Innovation and Value Creation. The inclusion of a higher number of levels may allow for a more detailed and specific evaluation, breaking down each dimension into finer stages of development. However, it may also increase model complexity and make practical application more challenging. A similar trend could be expected from models with 10 and 100 levels.

The matrix shows that the most critical dimensions for organizational maturity, such as Technology and Digitalization, Processes and Operations, Management and Strategy, and People and Competencies, are consistently present across models with different numbers of levels. Five-level models appear to be the most balanced and comprehensive in terms of dimension inclusion and frequency, suggesting that this structure is widely adopted for its effectiveness in capturing essential aspects of maturity.

The relationship between the number of levels and dimensions reveals that a higher number of levels may be associated with a more detailed evaluation of certain dimensions, while models with fewer levels may group or simplify aspects to facilitate application. The variability in dimensions and levels highlights the importance of selecting or designing a maturity model that fits an organization’s specific needs and emphasizes the most relevant areas for its context and strategic objectives.

4.3.2. Exploratory Inferential Analysis of Level Structure and Dimensional Breadth

To complement the descriptive results, we conducted an exploratory inferential analysis of the relationship between the number of maturity levels and the breadth of normalized dimensions captured by each model. For this purpose, we used the count of standardized dimension groups (n_dim_norm) as the dependent variable. As explained in Section 2.2, this count does not represent the raw number of dimensions originally declared by each author, but the number of thematic groups after normalization and aggregation (e.g., several author-specific labels may map onto a single standardized dimension such as People & Competencies or Technology and Digitalization).

On the side of maturity levels, we first considered only those models that define an explicit discrete maturity structure. The corpus for this analysis thus comprised 69 models with discrete levels (3, 4, 5, 6, 10, 11 or percentage scales), excluding models that do not formalize levels or that rely solely on continuous scores without thresholds. For the Kruskal–Wallis test, we grouped models into three categories: 3–4 levels (14 models), 5 levels (37 models) and 6+ levels (18 models, i.e., six or more levels, including the most granular schemes). The outcome variable was n_dim_norm, treated as an ordinal/non-normal count.

Descriptively, the mean number of standardized dimensions increased slightly with the number of levels: models with 3–4 levels had on average 2.43 dimension groups ($SD\approx 1.28$), models with 5 levels 2.68 groups ($SD\approx 1.49$), and models with 6 or more levels 3.00 groups ($SD\approx 1.41$). However, the Kruskal–Wallis test did not detect statistically significant differences between groups ($H=1.12$, $df=2$, $p=0.57$). In other words, while models with more levels tend, on average, to cover a slightly larger number of normalized dimension groups, this pattern is weak and not statistically robust within our sample.

As a complementary analysis, we treated the original number of maturity levels as an ordinal variable and computed Spearman’s rank correlation between the number of levels and n_dim_norm, using all models with a numeric level count (including those with 10, 11, or 100 levels and excluding only those with undefined levels). The resulting Spearman coefficient was low and non-significant ($\rho =0.10$, $p=0.41$), indicating no evidence of a systematic monotonic association between the number of levels and the breadth of standardized dimensions.

Overall, these exploratory inferential results are consistent with the descriptive patterns reported above. Five-level schemes remain the dominant design choice and are used both in relatively compact models (with few standardized dimensions) and in more elaborate ones. Models with six or more levels show a mild tendency to integrate a slightly broader set of normalized dimensions, but the effect size is small and statistically uncertain. Taken together, the findings support our substantive interpretation: the choice of maturity-level granularity (three, four, five, six or more levels) is not strongly constrained by the number of conceptual domains that the model aims to cover. Rather, level structure appears to be guided by design traditions, usability considerations, and sector-specific needs, while dimensional breadth is shaped more directly by the underlying conceptual scope (e.g., technology-only vs. techno-organizational vs. techno-socio-environmental models) than by the exact number of levels.

4.4. RQ.4 What Are the Most Frequently Incorporated Enabling Technologies in Industry 4.0 and 5.0 Maturity Models?

Another relevant analysis (Figure 7) emerging from the study of the 75 identified models concerns the enabling technologies of Industry 4.0 and 5.0. The key technologies driving digital transformation and the transition to new industrial paradigms exhibit differentiated adoption patterns between the two industries, reflecting their distinct objectives and approaches.

Industry 4.0 Models: Focus on Connectivity and Data

In the context of Industry 4.0, the most represented technologies in maturity models focus on process optimization, automation, and the interconnectivity of industrial systems. The Internet of Things (IoT) stands out significantly, integrated into 62 models, highlighting its fundamental role in real-time monitoring and remote management of equipment and production processes. Big Data ranks as the second most adopted technology, appearing in 55 models, underscoring the importance of collecting, analyzing, and leveraging large volumes of data for informed decision-making and continuous improvement.

Technologies associated with Automation, present in 49 models, form another pillar of industrial evolution, aimed at increasing operational efficiency and reducing human errors. Additionally, the use of cyber–physical systems (CPS), appearing in 31 models, reinforces the interconnection between physical and digital systems, enabling more precise control over industrial processes. Several analyzed technologies (such as IoT sensors, actuators, etc.) can be part of cyber–physical systems. Another key technology is Artificial Intelligence (AI), found in 19 models, emphasizing its capability to facilitate autonomous decision-making and enhance product and service customization. Technologies such as Digital Twins (13 models) and Cloud Computing (12 models) are also represented in Industry 4.0, enabling greater flexibility, scalability, and advanced simulation of industrial processes.

Industry 5.0 Models: Focus on Collaboration and Resilience

Industry 5.0, unlike its predecessor, places greater emphasis on human–machine collaboration and sustainability. Although its technological adoption is still limited compared to Industry 4.0, signs of steady growth are evident. The most representative technologies include Big Data, which remains present in four specific Industry 5.0 models, demonstrating that large-scale data analysis remains relevant but within a more collaborative environment.

One of the distinctive aspects of Industry 5.0 is the emergence of human–machine collaboration, present in one model, a technology that emphasizes the integration of human capabilities with automated systems to enhance process customization and adaptability. Automation remains important, albeit to a lesser extent, appearing in one model, indicating a shift towards more interactive and collaborative systems rather than entirely autonomous processes.

Cybersecurity, present in one Industry 5.0 model, also gains relevance in this context, given the increasing interconnectivity and the growing need to protect data and systems in increasingly complex environments. Other technologies, such as Collaborative Robotics and Edge Computing, have also begun integrating into Industry 5.0 models, each appearing in one model. These technologies enable greater responsiveness and process customization, key characteristics of this new industrial era.

Note: The lower number of Industry 5.0 models means that each listed technology appears in only one or a few models. Additionally, foundational technologies such as IoT and CPS, omnipresent in Industry 4.0, are not explicitly listed in the few Industry 5.0 models analyzed, although they are assumed to be part of the common digital environment.

Transition Between Industry 4.0 and 5.0 Enabling Technologies

Some enabling technologies show a smooth transition between Industry 4.0 and 5.0, suggesting their fundamental role in both industrial revolutions. The Internet of Things (IoT), present in two models shared between both industries, is a key technology for device interconnectivity, both in automated production environments and those requiring greater human interaction. Similarly, Big Data and Automation, each shared in one model, remain essential, although in Industry 5.0, they are expected to evolve toward a more collaborative and sustainability-focused approach.

Evolution of Enabling Technologies (2020–2024)

The period 2020–2024 revealed an evolution in focus toward different technologies, marked by the progressive emergence of the Industry 5.0 concept. In the early years of the period, maturity model publications were dominated by Industry 4.0 and its associated technologies. In fact, 2021 was the peak year with 18 new maturity models focused on Industry 4.0, followed by 2022 with 17 models. These publications reinforced the widespread use of IoT, Big Data, automation, and other technologies as pillars of digital maturity assessments during those years.

From 2023 onwards, a shift in trend is observed: although the number of new Industry 4.0 models declined (8 models in 2023), models related to Industry 5.0 emerged for the first time. In 2023, one exclusive Industry 5.0 model and two hybrid models (combining Industry 4.0 and 5.0 concepts) were published, marking the formal introduction of 5.0 technologies and approaches in maturity models. By 2024, interest in Industry 5.0 modestly increased, with two additional Industry 5.0 models published, while Industry 4.0 models continued to appear (13 in 2024). This suggests a gradual transition in the academic and industrial communities toward the new generation. The greater presence of Industry 5.0 models in 2023–2024 coincides with the incorporation of emerging 5.0 technologies (e.g., human–robot collaboration, edge computing) into maturity assessments.

5. Discussion

5.1. RQ.1 How Have Industry 4.0 and 5.0 Maturity Models Evolved Between 2020 and 2024?

Between 2020 and 2024, maturity models remained predominantly oriented toward Industry 4.0, representing approximately 92% of the total, with a peak in 2021–2022. However, a small but growing subset began to combine I4.0 with new dimensions or were explicitly identified as I5.0 models (the first hybrids and one exclusive I5.0 model appeared in 2023, followed by a gradual increase in 2024). This pattern suggests a gradual, partial, and uneven transition: human-centricity, sustainability, and resilience criteria are being added to the digital backbone of I4.0 (IoT, CPS, data), rather than replacing it. This interpretation aligns with other authors’ analyses [44,45,46,47].

Hybrid models are defined as those that (i) are identified as I4.0 but incorporate at least one I5.0 pillar (human-centricity, sustainability, resilience) in their dimensions or indicators, or (ii) explicitly define themselves as transitional I4.0/5.0 models. This operational rule guided the coding process and can be illustrated by MM.08 [48] and MM.18 [49].

In our corpus, purely I4.0 models declined from 17 in 2022 to 8 in 2023 (−47%), while hybrid models emerged in 2023 (N = 2). In 2024, I5.0-labeled models increased from 1 to 4. See the annual distribution in Figure 5.

To move beyond a merely descriptive synthesis of publication dates and thematic foci, we interpret the empirical patterns through an explanatory framework that links them to three interrelated mechanisms. First, the marked increase in Industry 4.0 maturity models from 2018 onwards can be understood as a response to the consolidation of Industry 4.0 as a dominant reference frame in manufacturing and operations research. Foundational reviews on Industry 4.0 technologies and applications have shaped a shared vocabulary and a digital transformation agenda [3,4,5,6,7,8], which in turn has generated demand for operational instruments capable of translating these broad visions into assessable capabilities. Readiness and maturity models are a natural candidate for this role, insofar as they offer a structured roadmap from a current to a target state (Teichert [29], Santos-Neto and Costa [35]), which helps to explain their rapid proliferation in domains such as SMEs, logistics, or digital twins.

Second, the selective emergence of models labelled as Industry 5.0 or hybrid 4.0/5.0 from 2021 onwards is consistent with the reorientation of industrial policy and scientific agendas towards more human-centric, sustainable and resilient production. Policy documents and academic contributions emphasise that Industry 5.0 builds on—rather than replaces—the technological foundations of Industry 4.0, but reorients priorities towards social and environmental objectives [21,22,25,38,44,45,50]. From this perspective, the rise of maturity models associated with Industry 5.0 reflects the efforts of specific epistemic communities to realign existing assessment schemes with these new priorities, for instance, by incorporating dimensions related to workers’ well-being, green transitions, or value creation for society.

Third, the analysed corpus exhibits a diffusion pattern that is characteristic of maturity model research more generally. Previous reviews show that once a field recognises maturity models as a legitimate instrument, design templates, sets of dimensions and level structures circulate across communities and sectors, giving rise to families of closely related models with incremental adaptations [32,34,51]. The concentration of models around certain areas of application (for example, SMEs, logistics, or employee competences) can therefore be interpreted not only as a temporal coincidence, but also as the outcome of knowledge transfer processes and the reuse of consolidated modelling patterns, mediated by conferences, special issues and national funding programmes. This three-mechanism framework helps to explain why the observed patterns display a defined structure and path dependence, rather than simply reflecting temporal co-occurrences.

This trend must be analyzed through a socio-technical lens, triangulating (a) policy and societal pressures (emphasizing sustainability, resilience, and human-centricity) that are reframing the governance of digital transformation beyond I4.0; (b) lessons from exogenous shocks that revealed resilience as a design requirement; and (c) organizational learning on human–machine collaboration. These pressures were translated into maturity models through (1) reweighting of evaluation criteria (beyond efficiency/productivity), (2) incorporation of new dimensions/indicators (e.g., worker well-being, AI/data ethics, circularity), and (3) expansion of validation contexts (e.g., disruption scenarios, safety/ergonomics).

To translate this conceptual shift into comparable assessment instruments, we operationalized the three pillars of Industry 5.0, human-centricity, sustainability, and resilience, in alignment with the literature that highlights the techno-centric limitations of I4.0 and the need to reorient objectives toward human and environmental goals [44,45,46,47].

• Human-Centricity. Maturity assessment should capture the quality of human–machine collaboration, control and decision-making distribution, work design and ergonomics, along with physical and psychological safety, privacy and data protection for employees, and active participation of workers in co-design and continuous improvement. This focus shifts away from the techno-centric logic of I4.0 toward a framework where workers are central actors in value creation.
• Sustainability. Maturity should reflect the integration of environmental principles into strategy and processes, including eco-design, resource efficiency, circularity, and climate impact management across the lifecycle, with traceability and transparency throughout the value chain. This shift aligns operational performance with planetary boundaries and growing societal and regulatory expectations [45,46].
• Resilience. In response to shocks and disruptions, maturity must incorporate the capabilities of anticipation, absorption, adaptation, and recovery, both in operations and in supply chains and digital systems. This includes organizational preparedness, business continuity, cybersecurity, process flexibility, and post-event learning as design components, not merely tactical responses. Thus, resilience, previously peripheral in I4.0, becomes a structural priority in I5.0.

This operationalization enables maturity models to be reweighted beyond productivity and offers transparent criteria for sectoral adoption, laying the groundwork for the implications detailed below.

For future maturity model design, the findings point toward explicitly multipillar architectures: retaining the digital core of I4.0, while integrating indicators for human-centricity, sustainability, and resilience with transparent weighting and scenario-based validation (e.g., supply chain disruptions, safety/ergonomics). For user organizations, a two-phase roadmap is recommended: first, consolidating digital capabilities and then deploying the I5.0 pillars through co-design and ethics-by-design approaches. Future research should delve into measurement validity (across sectors/regions) and explore trade-offs (e.g., productivity vs. well-being vs. environmental impact).

5.2. RQ.2. Are There Significant Differences in the Evaluated Dimensions Between Industry 4.0 and 5.0 Maturity Models?

There are differences, but not in the form of a complete replacement of I4.0 by I5.0; rather, they represent a reorientation and reweighting. I5.0 models add emphasis on human-centricity, sustainability, and resilience, while retaining the digital foundation of I4.0 (technology, processes, management). The result is a coexistence with relative shifts in priority.

In I4.0 models, the dominant triad is Technology and Digitalization–Processes and Operations–Management and Strategy. This reflects a design oriented toward productivity, efficiency, and standardization, typical of digitalization and automation. In contrast, I5.0 models shift priorities toward Sustainability and Social Responsibility and Human-–Machine Interaction (HMI), with the emergence and strengthening of People and Competencies as a transversal axis. This does not replace the digital core; it reinterprets it: technology remains, but becomes subordinate to human-centered and environmental goals.

Certain dimensions exhibit mixed behavior. Innovation and Value Creation and Infrastructure/Ecosystem appear in both approaches, but their narratives differ: in I4.0, they function as levers for efficiency and scalability; in I5.0, as enablers of sustainable, collaborative, and resilient business models. Other minor dimensions, such as Quality, Customers and Market, and Security and Governance, remain at low or irregular levels over the period. This suggests that many models either take certain frameworks for granted (e.g., quality management) or embed them implicitly within processes or technology.

Four driving forces explain the observed shifts:

• Regulatory and societal pressure, particularly after 2021, pushing for the explicit integration of sustainability, well-being, and ethics in maturity assessments.
• Exogenous shocks, which prioritized operational continuity, supply chain robustness, and cyber-resilience, elevating resilience from a “tacit requirement” to a design dimension.
• Learning about human–machine collaboration, which shifted automation “for” humans to automation “with” humans—impacting competencies, ergonomics, safety, and data governance.
• Path dependency, where the continuity of I4.0’s digital core explains why Technology and Processes remain central even in I5.0 models: I5.0 redirects objectives, but does not dismantle capabilities.

In terms of validity, the observed contrasts may be influenced by sample and taxonomy effects, since the normalization and mapping of terms affect the recorded presence/absence of dimensions. Temporal distribution also matters, as the emergence of I5.0 models in 2023–2024 concentrates new dimensions in a few models, potentially biasing annual comparisons. Lastly, publication bias plays a role: generalist models tend to cover more dimensions, whereas sector-specific models narrow their scope, though this does not imply lower maturity.

Based on dimensional analysis, future model designs should adopt a multipillar architecture that preserves the I4.0 core (technology, processes, management) while integrating, with explicit reweighting, the I5.0 pillars (human-centricity, sustainability, resilience). Terminological consistency should be maintained to avoid overlaps (e.g., distinguishing Security/Governance from operational Cybersecurity); validation should be scenario-based (e.g., operational continuity, supply chain, cybersecurity, human–machine collaboration); and context-specific adaptation by sector and organization size (SMEs vs. large firms) should be applied without losing comparability.

From a user perspective, a two-phase path is recommended: first, consolidate I4.0 digital capabilities; then, build upon them to deploy I5.0 pillars, prioritizing people, the planet, and system robustness. Privacy, safety, and worker participation should be treated as design and governance requirements, and the dimensional profile should be used to balance efficiency with socio-environmental value and business resilience, adjusting implementation to the sector and organizational scale.

The content differences (dimensions) between I4.0 and I5.0 are complemented by structural differences (maturity level configuration). In the next section (RQ3), we analyze whether maturity level structures relate to the selection and prioritization of these dimensions, and what this implies for the design of maturity models.

5.3. RQ.3 Is There a Significant Relationship Between the Number of Levels in Industry 4.0 and 5.0 Maturity Models and the Evaluated Dimensions?

The maturity level structure across models reveals a consistent preference for five-level discrete schemes during the 2020–2024 period. This configuration is interpreted within the corpus as an operational balance between granularity and simplicity. Alternatives exist, such as three-, four-, and six-level schemes, as well as outliers (10, 11, 100 levels or continuous/“no fixed level” scales), but these remain in the minority. Among the 75 models analyzed, 68 specify maturity levels, while 7 do not define fixed levels. The five-level scheme is the most prevalent (37 models; 49.33%), followed by six levels (14; 18.67%), four levels (8; 10.67%), and three levels (6; 8.00%). Designs with 10, 11, or 100 levels are rare (1 each; 1.33%). This preference for five levels remains stable throughout 2020–2024.

Descriptively, models with more levels tend to incorporate, on average, a slightly broader set of standardized dimensions, but the pattern is weak. When models were grouped into three categories (3–4 levels, 5 levels, and 6+ levels), the mean number of normalized dimension groups (n_dim_norm) increased modestly from 2.43 (3–4 levels) to 2.68 (5 levels) and 3.00 (6+ levels). However, an exploratory Kruskal–Wallis test did not detect statistically significant differences between these groups ($H=1.12$, $df=2$, $p=0.57$). Similarly, when the original number of maturity levels was treated as an ordinal variable (including 3, 4, 5, 6, 10, 11 and 100 levels) and correlated with n_dim_norm, the Spearman coefficient was low and non-significant ($\rho =0.10$, $p=0.41$). Taken together, these non-parametric tests suggest that there is no statistically robust monotonic association between level count and the breadth of standardized dimensions in our sample.

These inferential results qualify and refine the descriptive interpretation. Five-level schemes remain the dominant design choice and are used both in relatively compact models (with few normalized dimensions) and in more elaborate ones. Models with six or more levels show a mild tendency to cover a broader set of conceptual domains, but the effect size is small and statistically uncertain. This supports the substantive conclusion that the choice of maturity-level granularity (three, four, five, six or more levels) is not tightly constrained by the number of conceptual domains that the model aims to cover.

Instead, the evidence is more consistent with a design logic in which level structure is driven by cognitive and practical considerations (e.g., the ease of communicating and interpreting a five-step progression), by path dependence on earlier reference models, and by sector-specific conventions, rather than by a strict requirement to allocate one level per dimension or to systematically increase dimensional breadth with each additional level. In contrast, dimensional breadth appears to be shaped more directly by the underlying conceptual scope of the model (technology-only vs. techno-organizational vs. techno-socio-environmental) than by the exact number of maturity levels.

The few outlier designs with very granular or continuous scales (10, 11 or 100 levels) illustrate this point. They tend to be associated with highly specialized purposes (e.g., detailed monitoring of digital twin capabilities or fine-grained progress tracking in specific sectors) and require higher data and interpretation effort, but they do not necessarily span more standardized dimensions than some five- or six-level models. Likewise, models that do not define fixed levels typically prioritize qualitative diagnosis, gap identification, or prioritization of interventions over comparative scoring. In this sense, the inferential analysis reinforces a key practical implication: organizations and designers can choose between three-, four-, five-, six-level or continuous schemes primarily on the basis of usability, context, and communication needs, without strong evidence that one particular level count is intrinsically associated with broader or narrower conceptual coverage.

5.4. RQ.4 What Are the Most Frequently Incorporated Enabling Technologies in Industry 4.0 and 5.0 Maturity Models?

I4.0 models prioritize the connectivity–data–automation stack (IoT/CPS, analytics/AI, cloud/edge computing, digital twin, robotics/automation), while I5.0 models introduce stronger emphasis on technologies oriented toward human–machine collaboration and personalization, while retaining the inherited digital foundation. In the corpus, for example, automation-related technologies are present in 49 models, and cybersecurity is explicitly mentioned in only one I5.0 model. This supports the conclusion that the I4.0 focus remains dominant, although early signs of a 5.0 reorientation are emerging.

I4.0 is characterized by a logic of interconnection, data capture, and exploitation for decision-making and continuous improvement, combined with automation and advanced simulation. In I5.0, attention increasingly shifts to collaborative systems (e.g., HMI/cobots) and to the personalization of products and services; cybersecurity is also mentioned, although its presence is still limited. Several technologies, such as IoT and CPS, show continuity between I4.0 and I5.0, acting as a shared digital foundation.

The lower number of I5.0 models means that each occurrence carries greater relative weight. Furthermore, part of the digital infrastructure (e.g., IoT/CPS) may be assumed as a shared environment and is not always explicitly mentioned in the texts. This suggests caution when comparing frequencies across categories.

This analysis aligns with the findings of Hein-Pensel et al. [34] and Ünlü et al. [51], who noted that I4.0 maturity models often lack a human-centric focus, standardization, and practical validation. However, our contribution is both longitudinal and explanatory: (i) We quantify and interpret the 2020–2024 transition (I4.0 predominance, emergence of hybrids in 2023, and growth of I5.0 in 2024); (ii) we introduce an operational definition of “hybrid” that enables reproducible comparisons; (iii) we provide a meta-typology of maturity levels, highlighting the stable dominance of the five-level scheme and its rationale for comparability; and (iv) we present a technology cartography grounded in evidence-based coding.

5.5. Underlying Mechanisms: Policy Agendas, Epistemic Communities, and Diffusion

The patterns observed in our corpus, the delayed but then accelerating emergence of Industry 5.0 labels, the reweighting of dimensions toward sustainability and human-centricity, and the selective extension of technology stacks, are not only internal to the maturity-model literature. They also reflect broader political and socio-technical agendas that shape how Industry 4.0 and Industry 5.0 are framed in policy documents, strategic roadmaps, and academic debates. In this sense, maturity models can be read as boundary instruments that translate high-level industrial strategies into operational diagnostics and roadmaps at the firm level.

First, the rise of Industry 5.0 maturity models is closely linked to policy agenda-setting at supranational and national levels. The European Commission explicitly framed Industry 5.0 as a policy-oriented complement to Industry 4.0, aiming to reorient digitalization toward human-centricity, sustainability, and resilience rather than pure productivity gains [38]. Subsequent conceptual and review articles show how this narrative has been taken up and elaborated in the academic literature, sometimes as an alternative paradigm, sometimes as a corrective layer on top of Industry 4.0 [44,45,46,47]. Our finding that most recent models remain techno-centric but gradually incorporate sustainability and people-related dimensions is consistent with this trajectory: designers of maturity models appear to respond to evolving policy expectations by adding or reweighting dimensions that make these pillars assessable, while retaining the digital backbone of Industry 4.0.

Second, national and regional industrial strategies contribute to the way maturity is conceptualized and measured. Comparative analyses of industrial policy frameworks, such as Germany’s Industrie 4.0, the EU’s Industry 5.0, Japan’s Society 5.0, and Chinese strategies for intelligent manufacturing, show that they embed distinct priorities and imaginaries about the role of technology, the state, and society in industrial transformation. Bibliometric and conceptual studies of Industry 5.0 emphasize, in particular, the way European debates seek to align digital transformation with human-centric, sustainable, and resilient futures [50]. In parallel, Morillo Trujillo [52] shows how the strategies of China, Japan, and the European Union differ in their emphasis on innovation, competitiveness, social well-being, and geopolitical positioning. It is therefore unsurprising that many maturity models in our corpus adopt dimensions, labels, and examples that resonate with the policy context in which they were developed. From this perspective, maturity models do not simply measure a neutral “state of Industry 4.0/5.0,” but enact particular visions of what a desirable industrial future should look like.

Third, epistemic communities and diffusion processes help explain why certain level structures and dimensional configurations become dominant. Both our review and previous systematic mappings indicate that a relatively small set of reference frameworks exerts an outsized influence on subsequent designs, either through direct adaptation or by serving as conceptual templates [34,51]. Over time, communities of researchers, consultants, and policy experts stabilize a shared repertoire of dimensions (e.g., technology, processes, people, strategy) and level patterns (in particular five-level schemes), which are then replicated and customized across sectors and regions. Recent bibliometric analyses of Industry 5.0 highlight how these discourses spread internationally through publications, conferences, and research networks, while being locally translated and hybridized [21,50]. Our evidence of emerging hybrid models and selective integration of Industry 5.0 pillars fits this view: maturity models act as vehicles through which policy ideas, academic concepts, and managerial fashions are operationalized and circulated.

Taken together, these mechanisms imply that maturity models should be interpreted not only as technical assessment tools, but also as embedded artefacts of broader political and epistemic projects. For designers of new models, this calls for more explicit reflection on which policy narratives and value commitments are being encoded into level definitions, dimensions, and indicators (e.g., the relative weight of productivity, worker well-being, and environmental impact). For users and decision-makers, it highlights the need to critically examine the origin, assumptions, and implicit priorities of any given model before adopting it as a roadmap for Industry 4.0/5.0 transformation.

6. Limitations and Future Work

First, by design, our analysis is restricted to the conceptual and design features of maturity models as documented in the peer-reviewed academic literature. We do not examine how these models are adopted, implemented, or perform in real organizations, nor do we evaluate their impact on operational, human, or environmental outcomes. Such an assessment would require different types of data (e.g., longitudinal case studies, survey-based adoption and usage data, or usage analytics) and specific methodological frameworks, and therefore lies outside the scope of this review. Our focus on published model descriptions also means that we abstract away from tacit modelling practices, consulting artefacts, and informal adaptations that may occur when models are implemented in practice. We explicitly view this mapping as a necessary precondition for, rather than a substitute for, future empirical research on the practical uptake, effectiveness, and potential unintended consequences of Industry 4.0/5.0 maturity models.

Second, a limitation concerns the construction of the document corpus. The search strategy relied on two major citation databases (Web of Science and Scopus), a single search string, English-language publications, and a 2020–2024 time window. While these design choices support transparency and replicability, they may also have led to the omission of relevant work published in other languages, indexed only in alternative databases, or appearing in grey literature (e.g., industrial consortia reports, consultancy frameworks, or government guidelines). In addition, because both databases returned more than 4000 records, we used the built-in relevance ranking functions to select the top 250 hits from each source before screening. Although we mitigated potential selection bias by combining two databases, applying broad subject filters, and complementing the search with backward citation chasing, it remains possible that relevant but low-ranked or differently framed maturity models were missed. Our temporal focus on 2020–2024 also implies that earlier influential models are only indirectly represented through the way they are cited or adapted in the more recent literature.

Third, several limitations are associated with the operationalization and coding of variables. The classification of models into Industry 4.0, hybrid, or Industry 5.0 scope is based on the evaluative content of instruments (dimensions, indicators, level descriptions) and thus involves judgement in borderline cases where human-centricity, sustainability, or resilience are mentioned in the narrative but only partially reflected in the scoring structure. Similarly, the normalization of heterogeneous author labels into a standardized taxonomy of dimensions and technologies necessarily simplifies and aggregates distinct nuances, and different grouping rules might have produced slightly different frequency patterns. Although the GPT-assisted extraction protocol helped to systematize data collection, all codings were ultimately verified and, where necessary, corrected by the authors; we did not, however, compute formal inter-coder reliability coefficients. As a result, some residual misclassification or ambiguity in the assignment of scope, dimensions, or technologies cannot be completely ruled out.

Fourth, the analytical choices entail their own constraints. The corpus comprises 75 models obtained through a purposive, literature-based sampling strategy and is therefore not statistically representative of all possible maturity models in industry. The descriptive statistics and non-parametric tests reported in Section 4 are interpreted as exploratory diagnostics rather than as confirmatory evidence of population-level relationships. In particular, the Kruskal–Wallis and Spearman analyses of the association between level count and dimensional breadth are limited by modest group sizes and by the fact that models were not independently sampled with respect to these characteristics. Likewise, our study does not assess the psychometric properties, implementation costs, or empirical performance of the reviewed models; a simple frequency of occurrence of dimensions or technologies should not be equated with conceptual superiority or practical effectiveness.

Looking ahead, several avenues for future work arise from these limitations. To enhance transferability and comparability across studies, future versions of this mapping should incorporate quality and usage matrices, following approaches such as Ünlü et al. [51], including dimensions such as objectivity, tool support, and practical adoption. Additionally, sector-specific and SME-focused lenses, as explored in Hein-Pensel et al. [34], should be further developed in order to capture the contextual contingencies of maturity model design and use, including differences across regions, value-chain positions, and regulatory environments. Finally, maintaining and periodically updating the temporal series and the evidence-based rule framework that support our transition analysis will be essential to track how the conceptualization and practical role of Industry 4.0/5.0 maturity models evolve over time, particularly as more explicitly Industry 5.0-oriented or hybrid models are proposed and as empirical evidence on their use accumulates.

7. Conclusions

This study provides a longitudinal and explanatory view (2020–2024) of how maturity models (MMs) conceptualize the transition from Industry 4.0 to Industry 5.0. Empirically, we show that the field remains anchored in the digital backbone of Industry 4.0 (technology, processes, and management), while a smaller but growing subset of models progressively integrates human-centric, sustainability, and resilience-oriented criteria. To make this transition measurable, we operationalized the notion of a “hybrid” model at the I4.0–I5.0 interface and applied this rule consistently across a corpus of 75 models, rather than treating intermediate cases only in qualitative terms.

Structurally, we developed a meta-typology of maturity levels and confirmed the stable dominance of five-level schemes as a pragmatic balance between granularity and usability. Exploratory statistical diagnostics indicated that the number of maturity levels is only weakly related to the breadth of standardized dimensions, suggesting that level granularity is largely independent from conceptual scope. Content-wise, our results support an overlay interpretation: Industry 5.0 does not replace Industry 4.0, but adds explicit priorities related to people, environmental and social performance, and robustness on top of an existing digital foundation.

These findings have two main practical implications. First, for the design of maturity models, they point towards multipillar architectures that retain the Industry 4.0 core (technology, processes, and management) while explicitly incorporating Industry 5.0 pillars (human-centricity, sustainability, and resilience) with transparent terminology, weighting schemes, and scenario-based validation (e.g., supply chain disruptions, safety and ergonomics, or cyber–physical incidents). Second, for user organizations, they support a staged roadmap in which digital capabilities are consolidated first, and then used as a platform to implement human–machine co-design, ethics-by-design, and explicit management of trade-offs between productivity, worker well-being, and environmental impact.

Several limitations qualify our conclusions. The number of explicitly Industry 5.0 and hybrid models in the 2020–2024 period remains modest, taxonomic heterogeneity required normalization decisions that may influence some counts, and publication and database-relevance biases may have led to the omission of less visible models. Building on this work, future research should (i) integrate quality and use criteria (e.g., objectivity, tool support, and practical adoption) into comparative assessments; (ii) deepen sector- and region-specific analyses, especially for SMEs and developing contexts, while preserving comparability; (iii) strengthen empirical validation and cross-sector/cross-country measurement of maturity constructs; and (iv) examine how maturity profiles relate to performance under disruption, explicitly modeling trade-offs between operational efficiency, human-centric outcomes, and environmental sustainability.