A Comprehensive Framework for Integrating Extended Reality into Lifecycle-Based Construction Safety Management

Abstract

Construction remains one of the most hazardous industries, with high accident rates driven by insufficient planning, coordination, and safety training. While extended reality (XR) technologies, encompassing virtual, augmented, and mixed reality, have shown promise in improving safety outcomes, existing applications are typically isolated, lacking integration across the project lifecycle and alignment with digital methodologies such as those found in Construction 4.0. This study proposes a comprehensive workflow and framework for the integration of XR technologies into construction safety management, grounded in Building Information Modelling, Lean Construction, and Prevention through Design. This methodology structures the use of XR to support safety planning, training, inspection, and control, with a focus on lifecycle integration and proactive risk mitigation. Implementation examples are presented to illustrate the framework’s applicability and scalability. These demonstrate how XR can support immersive walkthroughs, synchronisation with BIM data, and simulation of human–machine interactions. This study contributes a structured, replicable approach that addresses the current fragmentation of XR safety applications, offering both a theoretical basis and practical guidance for adopting XR in construction safety workflows.

1. Introduction

The construction industry is widely recognized as one of the sectors with the highest rates of occupational accidents and fatalities worldwide [1,2]. Despite advances in regulations, safety protocols, and training programs, many incidents continue to be caused by factors such as lack of situational awareness, inadequate planning, and poor communication on construction sites. In response to these challenges, the adoption of emerging technologies has become a priority to improve safety performance and reduce risk exposure across all project stages [3,4,5].

In recent years, several emerging technologies have been explored to enhance construction safety management, including Building Information Modelling (BIM), Artificial Intelligence (AI), wearable devices, and Internet of Things (IoT) systems. These tools have contributed to proactive safety planning, real-time monitoring, and predictive risk assessment across various project phases. In this context, Extended Reality (XR)—which includes Virtual Reality (VR), Augmented Reality (AR), and Mixed Reality (MR)—has emerged as a promising technological paradigm to support safety management [6,7]. XR technologies enable immersive environments where users can experience simulated tasks, visualize risks, and interact with digital models in real time. These tools offer unique advantages for training, planning, inspection, and coordination, contributing to the development of a proactive safety culture [8].

However, these applications have generally been limited to isolated or short-term use cases, often disconnected from the broader project lifecycle or other digital tools commonly used in the Architecture, Engineering, Construction, and Operation (AECO) industry [5,7]. Moreover, many studies focus on the novelty of the technology itself, without fully addressing its integration into standard workflows or aligning it with methodologies such as Building Information Modelling (BIM), Lean Construction, or Prevention through Design (PtD) [4,6,9,10,11].

As a result, there is a lack of comprehensive frameworks that systematically incorporate XR technologies into construction safety management from a lifecycle and methodological perspective. Other technologies, such as Building Information Modelling (BIM) and Artificial Intelligence (AI), have partially addressed similar gaps by enabling proactive safety planning, predictive analytics, and real-time hazard detection. BIM-based frameworks facilitate lifecycle safety management through design-integrated hazard visualization and collaborative workflows, while AI enables automated risk assessment by analysing historical and sensor data. These approaches highlight the importance of structured integration for technological adoption in safety contexts. However, XR has not yet been fully incorporated into comparable frameworks, leaving its lifecycle application underdeveloped and its industry-wide potential untapped [12,13]. This gap hinders the potential of XR to become a practical and scalable solution for industry-wide adoption [4,14,15,16,17].

This article aimed to address this gap by proposing a structured framework for the integration of XR technologies into holistic construction safety management. The proposed approach is grounded in well-established methodologies (BIM, Lean, PtD) and aligned with the principles of Construction 4.0, providing a clear workflow and implementation strategy to deploy XR across different project stages. In addition, the study presents implementation examples that demonstrate the applicability and relevance of the framework in real or simulated construction contexts.

1.1. Construction Safety

The Architecture, Engineering, Construction and Operation (AECO) industry is considered one of the most hazardous in the world [1]. Its work dynamics, interactions between diverse professionals, and construction site characteristics that include outdoor work, unpredictable environments, dangerous machinery and tools and unique and varied projects, make safety management complex [2]. Thus, multiple factors affect construction safety, ranging from organisational concerns to job site characteristics, materials and equipment, in addition to factors associated with workers and work teams [3]. Despite the sector’s efforts to improve safety, its high accident rates have not decreased in recent years, and challenges for improving safety persist, though the goal is still ‘zero accidents’.

Today, the sector is evolving within the framework of Construction 4.0. This new way of working proposes the digitalisation of the work environment and incorporates emerging technologies that use robotisation and automation. Alongside this, the integration of management and collaboration methodologies, such as Lean Construction and Building Information Modelling, establish new frameworks that aim to reduce losses and increase continuous improvement and efficiency [4]. The aim is to integrate and connect physical work environments with virtual environments by using digital twins (DT) that make it possible to control and anticipate the behaviour of physical environments to improve the planning and execution of construction and operation processes and infrastructure (under development or during their useful life) [5].

Faced with the challenge of reducing the accident rate in the working environment as proposed by Construction 4.0, extended reality demonstrates multiple advantages in various aspects of safety management [6,7]. Various tools can be developed using virtual reality, which creates total immersion; augmented reality, the superimposition of virtual elements on reality; and mixed reality (MR), a combination of virtual reality and augmented reality that creates environments in which real and virtual elements interact [7,8,9]. The immersive possibilities of first-person BIM models, such as recreation of real environments, collection of site data and worker behaviour, overlap with reality and safe, virtual interactions for users with these elements offer potential improvements in occupational health and safety in construction [10,11].

Extended reality is relevant today as a new platform for human–machine interactions (HMIs) in the context of the new work dynamics proposed in Construction 4.0 [12,13]. Its great advantage, like other digitalisation tools, is prediction, the ability to anticipate real events and scenarios in controlled virtual environments. This allows us to model reality, anticipate problems and prevent them [14,15].

1.2. Deployment of Extended Reality in Construction Safety

Various authors have deployed extended reality for construction safety management. Developments have focused on training workers through virtual reality for identifying risks, working at heights and in hazardous environments and identifying collective and individual protection elements in the workplace [16,17,18]. In addition, various research projects have focused on training and using and handling construction machinery and equipment [19,20,21]. Workers’ behaviour in these virtual environments has also been studied, in training and facing workplace disturbances and environmental variables like noise, among others [22]. Research has also been carried out on design reviews and construction planning. In addition, augmented reality has been used to supervise safety environments in the workplace, to verify construction processes and the provision of protective elements [23,24]. Mixed reality has also been used to train workers to use complex tools and perform tasks [25,26,27].

Extended reality constructs immersive virtual scenarios, in which safety managers can efficiently recognise and study hazards [28]. Moreover, safety training that has been traditionally oriented towards passive teaching can now migrate towards active worker participation in XR environments [29,30]. Extended reality offers the possibility to recreate real scenarios in safe and monitored virtual environments. These scenarios increase safety awareness, motivate interest in learning and energise the training activity. The correct installation of health and safety measures on site and monitoring and controlling them throughout the project are significant tasks. All the scenarios and tools offered by XR can be superimposed onto the real site (which changes over time, depending on the type of tasks and projects), allowing for more effective supervision and control [31].

While XR experiences validate these varied uses, most of these experiences have been isolated or were only part of research efforts by specific working groups or companies developing ‘XR pills’ for specific training. At present, extended reality cannot be considered a mainstream tool in the industry [32]. Integrating a holistic view of XR into workflows, deploying the technology effectively, developing its functionalities and multivariate uses, testing it to scale and using it in real scenarios are necessary to spread its use and implementation [12].

1.3. Knowledge Gaps

In recent years, the application of extended reality, encompassing virtual reality, augmented reality, and mixed reality (MR), has gained increasing attention in the construction sector, particularly in relation to safety management. Numerous studies have demonstrated the benefits of XR technologies for worker training, hazard visualization, equipment identification, and simulated task performance. However, despite the growing interest and the development of isolated XR applications, the field remains fragmented. Most contributions focus on specific use cases or disconnected technological prototypes rather than on integrated systems that support safety holistically across the lifecycle of construction projects [33].

A central shortcoming in the current body of research is the lack of holistic and standardised frameworks that effectively embed XR technologies into comprehensive safety management processes [34]. While certain studies propose frameworks or workflows for XR deployment, these often fail to incorporate the broader methodological foundations needed for real-world implementation, such as integration with Building Information Modelling (BIM), Lean Construction principles, and Construction 4.0 environments. Moreover, there is a deficit in lifecycle thinking: XR is rarely used from the early design stages through to construction, operation, and monitoring. This limits the technology’s capacity to become a truly transformative tool for accident prevention [35,36].

In addition, few existing approaches explicitly address the multifactorial nature of construction safety. Current XR solutions tend to emphasize technical risks or spatial hazards while overlooking the integration of contextual factors such as environmental conditions, organizational culture, work dynamics, and human behaviour [37]. The lack of customisable and emotionally engaging XR experiences—which would align simulations with workers’ specific roles, experiences, and cognitive responses—represents another barrier to achieving deeper awareness and behavioural change. Although storytelling and gamification have shown promise in other sectors, their strategic use in construction safety remains underexplored and insufficiently justified [38].

Another significant gap concerns the technical-operational readiness and scalability of XR-based safety systems. Many studies stop at the conceptual or pilot level, without detailing the architecture required for implementation or the technical processes for data integration, interface design, and real-time updates. This creates a disconnect between academic innovation and industry adoption [39]. Moreover, the lack of standard validation procedures, benchmarking, and comparisons with traditional safety tools hinders the ability to assess XR’s real effectiveness. There is also limited reference to existing international standards, such as ISO 19650 Part 6 [40], which could guide the structured inclusion of safety data in digital environments [41].

1.4. Research Objectives

Having identified these problems and knowledge gaps, this research proposes a framework for an XR solution for holistic safety management in construction. Thus, this research sought to:

• Integrate a holistic and lifecycle-based approach to safety management, incorporating Prevention through Design (PtD) and considering the wide range of contextual, organizational, and human factors that affect construction safety across all project phases;
• Embed safety management within the digital transformation of the sector, through the structured integration of Building Information Modelling (BIM), Lean Construction workflows, and Construction 4.0 technologies, including digital twins and interoperable platforms that support real-time data use and decision-making;
• Position extended reality as a central paradigm for human–machine interactions, enabling immersive training, collaborative planning, and simulation-based analysis in virtual work environments aligned with emerging digital work cultures;
• Develop and operationalize a comprehensive XR framework, capable of integrating diverse models and data flows into virtual environments for safety-related tasks, providing a technically detailed, adaptable, and scalable methodology for deployment in real-world projects; and
• Demonstrate the feasibility and potential of the framework through detailed implementation examples, highlighting its practical applicability, capacity for customization, and potential for standardisation and future validation in field contexts.

The remainder of the article is structured as follows: Section 2 presents the research methodology and development stages of the study; Section 3 describes the core elements of the proposed framework, including methodological foundations, integration strategies, and key components; Section 4 discusses the practical applications, implications, and limitations of the framework; and Section 5 offers the main conclusions and outlines future lines of research related to usability, validation, and adaptation of XR solutions in construction safety.

2. Research Methodology

This research adopts a three-stage methodological approach aimed at consolidating and systematising current knowledge on XR applications for safety in construction and translating it into a structured and practical framework. This is a conceptual and methodological study focused on the design and theoretical validation of a framework. The methodology combines a systematic literature review, analytical categorisation of implementation requirements, and framework development supported by illustrative implementation examples. This approach was chosen to bridge the current fragmentation in XR-based safety tools and to ensure that the proposed solution is grounded in both academic research and practical considerations relevant to the Construction 4.0 context.

Figure 1 shows the research methodology used. It shows the stages, research tools, activities and deliverables of each activity. The methodology has three stages: Stage I: Background of XR frameworks for safety management in construction; Stage II: Requirements and considerations for the implementation of XR in safety management; and Stage III: Design of a proposed framework for an extended reality solution for holistic safety management in construction, and examples of implementation of this proposed framework. In Stage I, a state-of-the-art review was conducted to identify background information on frameworks or workflows that show how to implement extended reality for construction safety management. For this purpose, the Web of Science and Scopus databases were reviewed. The search strings included combinations of the following keywords: (“construction safety” OR “occupational safety” OR “hazards” OR “risk management”) AND (“virtual reality” OR “augmented reality” OR “mixed reality” OR “extended reality” OR “XR”). Publications between 2010 and 2025 (April) were reviewed. The inclusion criteria were: (a) studies addressing XR applications in the context of construction safety, and (b) publications proposing frameworks, workflows, or implementation strategies. Exclusion criteria included: (a) studies not related to construction or infrastructure projects, and (b) papers focusing solely on unrelated uses of XR (e.g., gaming, healthcare). A total of 32 articles met these inclusion criteria and were analysed in detail, and overall, 85 articles were included in the review.

The search in Stage II was based on the same search as in Stage I, and the requirements and considerations for the deployment of XR for integrated construction safety management were identified according to three main aspects: (a) a holistic vision of construction safety management; (b) approaches from Construction 4.0 and methodological trends in the sector; and (c) extended reality as a paradigm for human–machine interactions in new virtual work environments. Considering these requirements, a framework for an extended reality solution for holistic safety management in construction was proposed in Stage III. It considers the deployment of construction safety management tasks in XR environments linked to BIM models and the project life cycle in the construction sector. Additionally, examples of the implementation of the proposed framework are shown to demonstrate the proposal’s reliability. These illustrative examples serve to demonstrate the technical feasibility and internal consistency of the proposed XR framework in simulated scenarios. Although they do not involve data collection from real participants, they allow for a conceptual validation of the framework’s structure and potential applications.

The analytical categorization of the selected literature was carried out using a thematic coding approach, guided by the objectives defined for each stage of the methodology. For Stage I, studies were categorized based on whether they proposed frameworks, workflows, or case applications involving XR for construction safety. For Stage II, documents were analysed according to three predefined dimensions: (a) degree of holistic integration of safety management processes, (b) alignment with Construction 4.0 methodologies (e.g., BIM, Lean, Digital Twins), and (c) consideration of XR as a human–machine interface for immersive training and planning environments.

These analytical categories allowed the identification of strengths, gaps, and opportunities across the reviewed studies and informed the design principles of the proposed framework. In Stage III, illustrative examples were included to demonstrate how the framework could be applied in practice. These examples were selected based on their representativeness of key safety tasks and their relevance to different project phases (e.g., training, planning, inspection). The purpose of these examples is to validate the internal consistency and practical applicability of the framework, showcasing its flexibility and potential for integration into existing workflows.

Alignment of Research Objectives with the Manuscript Structure

This subsection presents a detailed alignment of the five research objectives and the specific sections and subsections of the manuscript in which they are addressed.

Table 1. Alignment of research objectives with the manuscript structure.

Research Objectives	Addressed in Section(s)
Integrate a holistic and lifecycle-based approach to safety management, incorporating Prevention through Design (PtD), and considering the wide range of contextual, organizational, and human factors that affect construction safety across all project phases.	Section 3.1 and Section 4.2
Embed safety management within the digital transformation of the sector through the structured integration of Building Information Modelling (BIM), Lean Construction workflows, and Construction 4.0 technologies, including digital twins and interoperable platforms that support real-time data use and decision-making.	Section 3.1, Section 3.2, Section 4.1 and Section 4.2
Position extended reality as a central paradigm for human–machine interactions, enabling immersive training, collaborative planning, and simulation-based analysis in virtual work environments aligned with emerging digital work cultures.	Section 3.1, Section 3.2, Section 4.1 and Section 4.2
Develop and operationalize a comprehensive XR framework, capable of integrating diverse models and data flows into virtual environments for safety-related tasks, providing a technically detailed, adaptable, and scalable methodology for deployment in real-world projects.	Section 3.2, Section 4.1 and Section 4.2
Demonstrate the feasibility and potential of the framework through detailed implementation examples, highlighting its practical applicability, capacity for customization, and potential for standardisation and future validation in field contexts.	Section 4.1, Section 4.2 and Section 4.3

mapping enables readers to clearly trace how each objective is developed across the conceptual, methodological, and applied components of the study.

3. Proposal of an Extended Reality Solution for Holistic Construction Safety Management

3.1. Requirements and Considerations

A holistic view of safety management in construction must rethink the processes, stages, methodologies and technologies that have been traditionally used. Different aspects must be considered for the deployment of the extended reality as a holistic solution for safety management throughout the project life cycle. In the following, different considerations and requirements are indicated to build the proposed workflow according to the research objectives.

3.1.1. Holistic Vision of Construction Safety Management

The traditional approach to risk identification and safety management in construction projects focuses on detecting risks and implementing mitigation measures once the design has already been defined. The safety officer joins the project when construction begins. By entering the project at this stage, risk prevention measures are reduced to incorporating protective elements (collectively or individually) and reducing the probability of incidents so as to conduct work more safely on the construction site [42]. Prevention through design (PtD) seeks to incorporate safety managers at an earlier stage of the project so that they can participate in the design and planning stages [43,44]. Thus, their role would not be reduced to incorporating protective elements and measures, but could also have an impact on project design and planning, making the construction process safer from the beginning [45].

Under this approach, the processes associated with safety management must be restructured. No longer would protective measures only be designed to reduce risks at the construction stage, but they would aim to influence design and project management decisions at an early stage. In addition to their involvement in the safety management planning and safety monitoring and control stages, safety officers will need to be involved in the design and construction planning stages [46]. Figure 2 shows the processes and activities to be considered in safety management.

During the design stage, design iterations and safety reviews should feed back into the typical project speciality designs. Construction planning should also incorporate safety data and collective protection measures. Construction logistics will be influenced by safety criteria that minimise the associated risks. When safety planning and integration are worked into the schedule, there will be a continuous review of safety plans, along with ongoing training processes [47]. Management methodologies and technologies can be used to implement these processes (Lean Construction, BIM, applications associated with virtual design and construction, and other tools within Construction 4.0). Finally, safety monitoring and control will be developed throughout the process. All planned measures will be aligned with applicable international and national standards, according to the characteristics and conditions of the particular project (e.g., ISO 45001 Occupational Health and Safety Management Systems [48]). In addition to ISO 45001, it is important to consider comprehensive regulatory frameworks such as the UK Construction (Design and Management) Regulations 2015 (CDM 2015) and the EU Directive 92/57/EEC, which provide structured approaches to integrating health and safety throughout the project lifecycle. These regulations emphasise early risk identification, coordination between duty holders, and the proactive management of safety during both design and execution. The proposed framework aligns with these principles by incorporating Prevention through Design (PtD), early-stage design review, and coordination of safety-related information within immersive XR environments. XR technologies offer new opportunities to support compliance with these regulations by enabling clearer communication, collaborative planning, and virtual simulations that reinforce worker awareness and safety decision-making [49].

In addition, the ISO 19650-6:2023 standard, which focuses on the collaborative sharing and management of health and safety information using BIM, provides a valuable framework that aligns with the objectives of this study. By incorporating information requirements, data structures, and responsibilities related to safety into the digital model, ISO 19650-6 supports the development of consistent and structured workflows that can be integrated into XR environments. This facilitates proactive hazard communication and traceability throughout the project lifecycle.

Along with the processes and activities in safety management, it is important to understand and correctly manage the factors that affect safety in construction. In this way, it is possible to align and implement these activities according to these indicators and to make the measures compliant. This verifies compliance with regulations that focus on avoiding accidents in the project itself (often considered minimal) and achieves a comprehensive and holistic vision of safety management, transversal to all those involved, and with a complete vision of the organisation.

Muñoz La Rivera et al. [50] consolidated and updated various studies and proposed 100 factors influencing safety on construction projects (fSCPs). These were grouped into four broad categories: (A) general aspects, which include factors associated with general organisation, standards, finances and productivity, the culture and safety climate and elements associated with previous lessons for establishing safety management programmes and systems; (B) materials and equipment, which considers the condition and characteristics of the materials and equipment used with respect to their influence on accident potential; (C) construction site, which incorporates site conditions and how these, together with work processes, can promote or prevent the occurrence of accidents; and (D) human aspects—worker and work team, which are associated with workers’ competencies, interactions, communication and motivations, and individual and collective responsibility for safety.

Although the control of and compliance with 100 factors could be complex, Muñoz La Rivera et al. classified these factors according to the Construction Accident Causation Framework (CACF). This makes it possible to understand the relationships and prioritisation levels in the factors regarding accident occurrence according to whether there are shaping or immediate factors or originating influences in the causes of accidents. Teizer [51] also suggested there is a wide spectrum of actions in safety management, ranging from various policies and prevention training to actions that can be taken immediately before an accident occurs in real time. Considering these two views, Figure 3 shows conceptually the link of the fSCPs, as categorised in the CACF, with the spectrum of actions proposed by Teizer. In the central area of the figure, safety management actions can be associated with the fSCPs and thus have direct control over how these factors are addressed and fulfilled. The location of these safety management actions will be in accordance with Teizer’s feedback-level pyramids and Heinrich’s pyramid, depending on the frequency with which the data or measures are implemented or their intended impact, and the associated organisational level.

3.1.2. Approach from Construction 4.0 and Methodological Trends in the Sector

BIM is spreading rapidly in the construction sector. It has multiple benefits and is transversal to all project stakeholders. However, its mass deployment is not homogeneous in project typology, uses or dimensions, among other aspects. Despite this, the deployment of extended reality needs to use input and connect with BIM environments [52]. This interconnection must be technological and methodological. The technological connection seems obvious, aiming at the use of 3D/4D/5D/8D models (parametric three-dimensionality, time, cost, safety), and aiming to incorporate other dimensions [53,54]. But at the methodological level, the integration of people and processes is relevant to functionality in working with the new XR tools. In this sense, the Lean Construction philosophy must also be deployed in virtual environments [55]. Good management practices, waste reduction and continuous improvement of construction management processes will be enhanced with virtual environments and provide new tools for monitoring and control [56,57,58]. The incorporation of Lean Construction workflows responds to their relevance in digitalised project management and their strong alignment with safety objectives. Lean planning tools such as the Last Planner System and pull scheduling facilitate short-term planning and collaborative decision-making, which can be enhanced through XR environments by providing visual and immersive representations of planned tasks and their constraints. This synergy supports proactive safety management by making potential risks visible and aligning team actions with project realities in real time.

Construction 4.0 promotes the automation, digitisation and virtualisation of the construction sector and encourages collaboration, integration and process-efficiency methodologies, such as Lean Construction and BIM, along with technologies that interconnect systems associated with construction project management (sensors, additive manufacturing, Big Data, VR, among others) to be deployed and supported [59,60].

In response to the demand for automation and digitisation, the concept of digital twins has emerged in recent years [61]. The term refers to a virtual representation of a physical system that provides accurate remote monitoring and control [62]. It is associated with system modelling, simulation tools and synchronisation devices, which link elements of the real and virtual worlds through sensors. By providing a link between these worlds, they allow users to visualise changes based on real-time data [63], which is a key feature of DT. In the digital building twin (DBT), this capability is directly related to real-time monitoring, diagnosis and forecasting in infrastructure under construction and operation [64,65].

Figure 4 shows a simplification of the DBT concept, with a focus on the construction phase. In a DT of a building at the ‘n’ stage, the DBT(n) will receive real-time data inputs from the project based on the information provided by the DBT(n − 1), which has ensured that what has been built in the RB(n) (physical building) is correct. Therefore, DBT(n) will consist of the original model and the changes made up to that point. With this information and the captured data, users can process and correct the model according to its behaviour in real time during project execution. This information allows the DBT(n) to deliver reliable information to continue during the construction of the RB(n + 1). This process continues until the end of the project, where the digital building is equivalent to the constructed building and its high representativeness as a constructed model is obtained because it was generated before and after each real-time instant.

Although there have been technological advances in the sector in the last decade, gaps in the technological–methodological coupling and the transition to new working methods make it difficult to streamline these elements and identify the role of DTs in construction. Special attention must be paid to the human factor and multidisciplinary teams on construction sites when considering a transition to a more technological environment. Construction workers are among the least qualified and have one of the highest accident rates and an ageing workforce.

Within the methodological trends of Construction 4.0, it is also relevant to consider the integration of advanced planning and control tools such as the Line of Balance (LoB) method. LoB is especially effective in projects that involve repetitive activities and sequential workflows—common in housing developments, linear infrastructure, or modular construction. From a safety management perspective, LoB provides a clear visualisation of work progression across time and space, enabling the anticipation of congestion, task interference, and access conflicts. Although traditionally used as a production planning tool, LoB also holds great potential for reinforcing safety coordination when integrated with digital environments. Its incorporation into visual or immersive systems can support the early identification of spatio-temporal safety risks that emerge from misaligned workflows or overlapping crews. As the industry moves towards more interconnected, real-time, and simulation-based planning processes, integrating methods like LoB within the digital ecosystem of Construction 4.0 becomes increasingly relevant—not only for efficiency, but also for proactive and dynamic safety planning.

3.1.3. Extended Reality as a Paradigm of Human–Machine Interactions in New Virtual Work Environments

Today’s world aims to develop a strong connection between the real world and the virtual world. From the intelligence associated with the Internet of Things (IoT) and data processing to the deployment of actuators in the physical world, current (and future) developments aim to generate these digital twins and connect them to real environments. However, this interconnection is no longer only of ‘things’ but also people and their interactions. The goal for these interactions is not to remain in the third person (the internet and social networks, for example, already manage to generate virtual identities), but to become immersive environments that improve user interactions (that are more than just viewing a screen and using a keyboard or a mouse) [66].

In the search for this new form of interaction, extended reality is becoming the new paradigm for human–machine interactions [67,68]. Buildings will be constructed with machinery, robots, prefabricated and sensorised equipment and operated from virtual environments [13]. Therefore, the workers of the future will have to know how to control and monitor construction from these virtual environments. Traditional methods of designing, planning, executing, and monitoring construction projects will shift to a highly digitised workplace. Workers will have to migrate to these new project execution methods and develop the competencies needed in this new way of working (and interacting with the physical and virtual worlds) [69,70,71].

These new virtual, interactive environments will permeate all productive sectors. While the products of the construction sector are ‘real world’ physical structures, their connection to their digital twins; the data associated with this infrastructure; and their handling, processing and analysis must be carried out in virtual environments [72,73,74]. The benefits are clear, and although a ‘construction metaverse’ may be far in the future, there is certainty that sooner rather than later work will be conducted in these virtual environments [70]. In view of this, all extended reality technologies, with greater or lesser levels of interconnection, represent an important advance in the sector to ‘train’ for the revolution that working in these new environments will present, especially considering that the concept of digitisation in the industry began only a decade ago with BIM [75].

3.2. General Description of the Proposed Framework

Under the challenges and new contexts in the sector, and in response to the requirements and considerations identified, the proposed framework considers the methodological and technological trends of Construction 4.0 and Prevention through Design and generates the basis for the comprehensive deployment of safety under these new paradigms. The framework seeks to advance the automation of the development and use of customisable extended reality applications for the integrated management of safety in construction, incorporating four key aspects: (1) Immersive, augmented and mixed visualisation and interaction, deploying different technologies that are adaptable to different safety management needs; (2) Customised extended reality tools to adapt developments to the specific requirements and contexts of the different projects where they will be used; (3) Integrated elements for a holistic view of safety to incorporate and integrate all safety requirements, actors and roles, and information models throughout the project life cycle; and (4) Storytelling to develop XR experiences to transmit and generate real awareness of safety in construction [76,77,78,79,80].

Figure 5 shows a simplified way the zones and elements are used in the proposed framework. Four major zones are considered: data and model inputs; XR programming, modelling and developments; XR interaction environment; and physical domain. The processes in project development and safety management interact within these zones, from which groups of data and models are derived. These data and models are collected, integrated and combined to develop the required experiences and processes in extended reality. After that, the experiences are deployed in the XR interaction environments to realise the safety management processes and tasks. These processes and tasks are developed cyclically and fed back into the models and building and safety management plans to their deployment at the design level and their implementation at the organisational and workplace levels.

Considering the elements and the simplified operation described, Figure 6 shows the complete version of the framework for an extended reality solution for holistic safety management in construction. The framework incorporates four interaction zones, 12 model packages and six data packages. Some aspects overlap and serve as input in different tasks. It is important to note that this distribution should be considered a general recommendation. Depending on the deployment of this proposal in a specific project or organisation, model or data packages may not be used, may be grouped or broken down, or new information may be added as required. To develop XR experiences, some information can be used directly from the traditional project stages, some will have to be improved or adapted for use, and other elements will have to be created to satisfy specific requirements in this new form of XR interaction.

In practical terms, the framework in Figure 6 structures the models, information and elements to deploy the tasks associated with safety management in XR environments according to the four groups of tasks within the XR interaction environment. The following describes the different models and data shown in the framework and how they will be incorporated into XR development environments. Several implementation examples are shown to clarify the developments.

The four major zones are structured vertically in the framework. Each zone receives inputs from the previous zone, and in turn, delivers outputs in a cyclical work process. These delimit the work to be completed and divide the efforts of each sector. However, aspects may overlap or complement one another to strengthen models, information packages or processes. The roles of each work zone are described in the following sections.

• Data and model inputs (requirements). Firstly, this area considers the processes in the safety management of a construction project and the respective data and model packages associated with each of these processes. Different data and models will be incorporated into the processes associated with design development, construction planning, construction safety planning and safety monitoring and control. There will be BIM models of the project; safety analysis models; models and information associated with the terrain, construction elements and their planning; models and information on protection equipment, threats and accidents; along with work dynamics and information updates during project execution. The processes, models and data have been distributed within sub-zones according to the BIM dimensions, including BIM 3D (three-dimensional parametric models), BIM 4D/5D (schedule and costs) and BIM 8D (health and safety). All these elements will feed the XR environments. Although they are explained as inputs to the process, they enter into an iterative cycle. They will improve and will become more robust as a result of the different processes, cycles and iterations in the framework, receiving feedback and ultimately updating the project’s design and management according to the safety perspective.
• XR programming, modelling and developments. This area incorporates all the necessary aspects to develop the XR experiences, taking as inputs the models and data from the previous area. First, the data and models are reviewed to analyse whether elements need to be improved or optimised (e.g., BIM models may be too heavy to be supported in XR and/or contain more information than required, depending on the purpose of the XR implementation). Following this review, models should be improved and/or replaced for optimal use in XR (as the process matures in an organisation, optimisation should occur at the start of the project, and protocols should be implemented to facilitate this). So far, only BIM models and planning information are available (i.e., inputs for reconstructing XR scenarios; however, they are ‘lifeless’ scenarios). Therefore, new elements will be added to this information, focused on achieving interaction with the models and deploying the required actions and effects. Information will be added on construction site objects, workers in virtual or mixed environments (roles and bots), and aspects related to climate and other variables. For training, gamification and storytelling aspects will be required. All this will be integrated into XR development environments (usually, game development engines or specific commercial XR applications), including tasks associated with the physicalisation of models and environments; creation of people and interaction effects; effects for virtual scenarios, implementation of work dynamics and all required interactions; together with links to simulation models in different areas of interest. In parallel, according to the XR technology used, the developments will be packaged for subsequent interaction through glasses, computers and mobile devices, among others.
• XR interaction environment. This area involves the deployment of safety management tasks from XR environments, through virtual, augmented or mixed reality. The technology will depend on the definitions of the project. In the early stages, virtual reality will be used more often, and will then be combined with mixed or augmented reality for deployment once the project is under construction or operation. The deployment involves the following tasks, with a focus on project safety: safety design and safety review; safety data and model preparation; protection system and PEE selection; safety planning and scheduling integration; safety planning review and safety training; operation and construction; and daily inspections and records. All these tasks aim to define or redefine safety management models or definitions, from specific protocols within the project to protection elements, work dynamics and work site configurations, among others. This will generate information that will feed back into the models or project data in an iterative cycle. This exchange of information is bidirectional and tends to be automated (this will depend on the robustness of the platforms and development; therefore, in the early stages of implementation, rework and information cycles may be present, seeking different options for their integration).
• Physical domain. This zone pertains to the deployment in the physical construction site of the actions studied and analysed in the XR environments. This can occur through training, task execution, inspections or the performance of machinery or equipment that could eventually be interfaced with the XR environments. This deployment overlaps with the previous area, but its deployment is relevant especially when considering the use of XR tools directly on the job site (e.g., in an augmented reality inspection). Moreover, when thinking about other Construction 4.0 technologies and digital twins interacting via XR tools, this zone becomes relevant, pointing towards XR technologies as tools for human–machine interactions in the context of Construction 4.0. Although graphically the flow in the Figure 6 depicts the interactions between the physical domain and the XR interaction environment zone and the project management and model development, the processes of a construction project also involve interactions with the physical environment, which for simplicity (and given the focus on the XR safety environment) have not been shown in detail.

4. Operationalization of the XR Framework for Safety Management

4.1. Implementation Examples

The following sections present each of the components included in the proposed framework, along with examples of their practical implementation.

4.1.1. Rules and Guide/Organizational Safety Guidelines

Regardless of the tasks or processes to be implemented, it is important to indicate that as initial process data, applicable safety rules and guidelines must be considered, according to the country or legal provision required. Additionally, organisational safety guidelines should be applied. These guidelines will govern the safety management process. Extended reality environments are therefore intended to be a tool to deploy the standards more rigorously. To no extent will the use of XR supplant any regulatory requirement. On the other hand, the implementation of work protocols based on XR can modify the organizational safety guidelines from a protocol or process standpoint, but not from a legal standpoint.

4.1.2. BIM Models

First, as part of the design development process, 3D BIM models of the project are produced. These consider all the models required for the project, generally separated into three categories: architectural models, structural models and mechanical, electrical and plumbing (MEP) models. The models required will depend on what is to be implemented. For example, if safety management in concrete formwork and pouring processes is to be implemented, only the structural model may be required. The level of detail (LOD) of the BIM models will be associated with the levels of information required and the graphic quality necessary for XR environments. The first aspect will be associated with the functionalities of the XR applications and the second with the desired level of realism. In both cases, more information and realism can be provided in the XR development environments at a later stage, although this is not optimal as the ideal is to avoid rework.

Figure 7 shows an example of the development of a BIM model of a reinforced concrete building. This was modelled based on CAD drawings (the original project did not include BIM models). Thus, the different structural elements are shown and modelled as independent elements, with a LOD 300. Each element contains geometrical information, location data, characteristics and other data of interest for the project. Different BIM software exists on the market and is useful for the development of 3D models (Autodesk Revit, ArchiCad, Tekla, others). In this example, Autodesk Revit software (version 2020) was used.

BIM models serve as inputs to reconstruct simple XR environments (from the project model only) and to analyse, for example, whether the designs proposed by engineers and architects comply with safety considerations. Here, prevention through design is relevant because the safety review is performed in the early stages of the project and the design itself can influence risk reduction.

To perform this task, video game development engines can be used as development software for XR experiences. Unity 3D or Unreal Engine are the most commonly used, as well as other commercial software that facilitates the process of visualisation and interconnection, such as Unity Reflect (which connects Unity and Autodesk Revit). In this software, the developer will have to incorporate all the physics required for interaction with the models (typically, colliders and gravity, among others) with textures that give them realism. Figure 8 shows an example incorporating physics and textures into a BIM model (Autodesk Revit) in Unity 3D (version: 2019.4) software to develop a more realistic model for interactions. Specifically, the image shows how colliders have been applied to individual building components (e.g., walls, floors, equipment), enabling physical interactions within the XR simulation (e.g., collision detection and walking through the space). Additionally, gravity parameters have been configured to simulate realistic object behaviour in dynamic scenarios. Textures and materials have also been applied to enhance visual realism and improve immersion. These modifications are made directly in Unity’s Inspector window, where the developer can assign physics components, adjust material properties, and fine-tune the visual and functional characteristics of the imported model. This setup transforms a static BIM model into an interactive, navigable XR environment suitable for safety simulations, training, or inspection tasks.

4.1.3. Safety Analysis Models

A risk analysis must be performed. The risk or accident analysis for the project is reviewed, associated with the BIM model, along with construction stages or specific construction processes, as required. Figure 9 shows an example of the implementation of a risk analysis for the fall of workers at the different levels associated with the construction of reinforced concrete walls. In this case, colour coding has been implemented in Autodesk Revit to visualise the zones of higher probability or severity for this risk.

4.1.4. BIM Models and Safety Analysis Models in XR Environments

Once the BIM models are incorporated into the XR development environments (with improved textures and added physical ones), and the safety analysis models are in the BIM environments (or others that could be incorporated), it is now necessary to move towards integrating tools and developing functionalities in the XR environments to interact with the models themselves. To interact with the model and take a virtual tour, a ‘First Person Controller (FPC)’ is required. This will allow us to interact with the model in first person, to walk through it, touch it and interact with the logical physics of real (or not real, as required) behaviour, among other functions. Its development and details will depend on whether it is generated for use with a computer or tablet (non-immersive) or virtual reality glasses (total immersion). The interaction capability of this ‘first person’ experience will depend on the developer. For example, buttons can be implemented on each of the building elements to display the information associated with each construction element, or information can be added on the risks of each area of the building. Figure 10 shows an example of the development in Unity 3D of an FPC. Different cameras, viewpoints, levels or visualisation layers of the model can be programmed. Although the figure shows a character, the FPC may not consider it a character.

As more actions are implemented, development efforts will increase. However, based on BIM models, risk analysis and safety regulations (general or internal to the organisation), XR experiences can first be developed for safety design and safety review. Further functionalities and the integration and enhancement of the previously described tools are explained in the following sections.

4.1.5. Terrain and Physical Environment Models

As part of the construction planning process, but also as part of the design stage, ‘terrain and physical environment’ models consider the topography and surroundings of the construction site. Depending on the tools with which they were obtained, these can consist of contour lines with simple discrete points (a traditional topography) or more robust point cloud models obtained by drones using photogrammetry or laser scanners. These can be integrated directly into BIM environments (allowing integrated work in those environments in the initial design stages) or later in XR development environments. Considering the weight of this type of file, simple models (with less graphic quality) can be incorporated in the initial stages of the project or according to the requirements, and others with more detail can be used in stages where better graphic quality is required (for example, for the XR immersion). Figure 11 shows an example of a 3D environment reconstructed using photogrammetric techniques with images captured by drones. In this case, the model is used for construction planning. However, more simplified models (e.g., contour lines) will be required for the initial design.

To provide context to the building site within a virtual or mixed environment, models of the environment can be incorporated. Figure 12 shows how a city environment has been implemented in Unity 3D. To make the generated scenario more realistic, the virtual environment should be similar to the real one. Based on the topography, an environment can be generated with elements of nature (mountains, vegetation), urban elements, buildings and streets, among others. It is important to consider that these elements add weight to the simulation; therefore, their graphic quality and properties should consider their level of interaction. For example, nearby streets and pavements can be implemented with a high level of development (e.g., where a worker will enter the construction site) and distant elements can be empty or be meshes for simple visualisation. Elements may be modelled or incorporated from free or paid 3D city model databases.

4.1.6. Construction Support Objects

Construction support object models include all elements that do not belong to the main building but support its construction (e.g., formwork or construction props). These should be included in the BIM model or in later in XR environments (low LOD models could be used in BIM environments; however, they should be upgraded or replaced in XR environments with the necessary construction site objects). Figure 13 shows an example of a framework modelled in Autodesk Revit, which can then be included in XR environments.

4.1.7. Construction Site Planning

Construction site planning information considers construction planning and associated cost planning in BIM environments (BIM dimensions 5 and 6). Various BIM software allows the integration of these schedules, linking time and cost to each element in the 3D BIM models. For example, Figure 14 shows the integration of time and cost planning to the BIM model in Figure 7 in Autodesk Navisworks software (2019), where the Gantt chart has been linked. In BIM modelling environments, information associated with the timing of construction could also be included, such as through construction phases (various predefined time scenarios). However, linking to planning charts from other potential linked software is not possible.

In addition to planning the main building and its construction process, the construction site will also be planned. While zones are generally predefined and established for the duration of the entire construction process, they will change as construction progresses. Thus, as shown in Figure 15, construction site planning can be implemented in the XR development software (in this case, Unity 3D), placing elements and activities according to their spatio-temporal variation.

Figure 16 shows the integration of various elements in the XR development environment (in this case, Unity 3D). The BIM model of the building at stage ‘n’ of the construction process (Figure 14) has been added to the physical building environment shown in Figure 12 within the planned construction site according to Figure 15. It is possible to see the gradual integration of each element and the versatility to modify them according to the project’s progress and the objectives defined for the use of extended reality.

4.1.8. Collective Protective Equipment (CPE)

As part of the construction safety planning process, the models and information sets associated with safety are included within the BIM environments and subsequently within the XR environments. Collective protective equipment (CPE) models consider all BIM elements of collective protective equipment, which are included in the BIM models. Several considerations must be taken into account. The development of this model includes input and output from the process; therefore, the iterations proposed in the framework will result in a final model with all these elements. The assignment of these elements will be associated with the progress of the project according to the different stages of construction. It is recommended that the BIM elements of the CPEs are modelled or incorporated within these environments, because while they can be modelled and their LODs can be improved in the XR environments, safety management is based on the BIM model, and therefore it is the direct link in these environments (either to other software or management methods).

Figure 17 shows an example of CPE implementation in Revit and its transfer to the XR development environment Unity 3D. On the other hand, Figure 18 shows the incorporation of other CPE directly into the XR environment. In this case, scaffolding is incorporated along with safety information signs. 3D models of SSCs can be developed in these same environments or generated in other 3D modelling software.

4.1.9. Personal Protective Equipment (PPE)

Along with CPE, personal protective equipment (PPE) must be considered. Because this is used by workers within the construction site, it will be included in XR environments on bots or first-person avatars. While personal protection elements will be associated with different tasks or work zones (which could influence information, for example, from the BIM environments and models), they will be used directly by workers, so their major deployment will be in the XR environments. Figure 19 shows the layout of different PPE in a construction site hut. In the XR simulation, the user can choose, according to the work, which elements to use (for training purposes, for example). Figure 20 shows PPE on bots and users in virtual simulations, either for a specific job or as general protection elements—in this case, using a mask to prevent COVID-19. The versatility of XR tools allows them to be adapted to the particular requirements of each case. As with CPEs, PPE can be imported from other modelling software or generated in XR environments. While Figure 20 shows PPEs on bots.

4.1.10. Hazards and Accidents

XR environments require detailed information on accidents and hazards at the construction site so that these can be programmed and replicated in virtual environments. This information may be structured differently, depending on the implementation requirements. Thus, it will be possible to have information on the types of general accidents, either by zone, type of construction or work carried out. This information, compiled, for example, as descriptive cards, will help the programmers to create animations with the accidents, define “action trees” of the accident that occurred before the behaviour (correct or incorrect) of the users in the virtual environment, display accident information, and others.

The casuistry of implementing hazards and accidents in virtual environments is wide. For example, for training purposes, the information will instruct workers or make them “suffer” accidents in virtual environments before performing incorrect behaviours. In addition, the association of accidents with areas in the work environment will allow for measuring the degree of risk to which workers are subjected or the probability of the occurrence of accidents according to how the workplace is configured, among other uses.

The implementation of accidents in virtual environments is a challenging task. All the casuistry and “stages” of an accident must be considered to cover the possibilities to which a worker may be subjected. Therefore, beyond the accident (the fall itself, for example), the actions close to the accident will be useful to simulate and make the worker aware of the details to avoid the accident itself. Thus, all accident scenarios should be programmed in such a way as to cover these spectra, according to the interest of the safety officer and the objectives of the training or the use of XR environments.

4.1.11. Construction Activities/Work Dynamics

Extended reality environments require detailed information on accidents and hazards at the construction site so that these can be programmed and replicated in virtual environments. This information may be structured in different ways, depending on the implementation requirements. For example, general information may be available on types of accidents, either by zone, type of construction or work. This information, compiled, for example, as descriptive cards, will help the programmers to create animations of the accidents, define ‘action trees’ of accident occurrence in response to the behaviour (correct or incorrect) of the users in the virtual environment and display accident information.

The rationalisation for implementing hazards and accidents in virtual environments is broad. For example, for training purposes, the information instructs workers or makes them ‘suffer’ accidents in virtual environments when incorrect behaviours occur. But it can also help determine which areas of a construction site have higher degrees of risk or the probability of accident occurrence based on how the workplace is configured.

4.1.12. Workers (Roles and Bots)

The deployment of activities, risks and work dynamics can be realised differently within virtual environments. Depending on the objective (training, design, planning), these can be included as programming aspects at the time of the ‘virtual walkthrough’ or as aspects that define the behaviour of virtual workers or first-person users. There, the creation and development of these components should consider aspects such as graphical and animation realism, behaviour, cause–consequence relationships and interactions between bots and users, among others. Figure 21 illustrates the integration of a virtual worker (bot) into the Unity 3D environment to simulate construction site activities. The character model and its animations were sourced from Mixamo, a widely used platform that provides rigged 3D characters and motion sequences. In this case, the bot represents a construction worker performing specific tasks. Once imported into Unity, these animations can be triggered and controlled through scripts or timeline-based logic to simulate realistic work routines. The Inspector window in Unity allows for adjusting animation controllers, movement patterns, and interaction triggers, enabling a detailed and behaviour-driven simulation. These bots are essential for representing human activity in XR training scenarios, especially for simulating unsafe behaviours, risk exposure, or interactions with machinery and users in immersive walkthroughs.

Figure 22 shows the integration of construction machinery into the XR environment, specifically demonstrating how predefined movement paths can be programmed in Unity to simulate the operation of vehicles or equipment within the site. This allows for the observation and analysis of how users—whether real participants or bots—respond to dynamic elements such as moving cranes, trucks, or lifting equipment. Such simulations are valuable for evaluating user behaviour, identifying potential safety conflicts, and testing alternative spatial configurations or safety protocols. Figure 23 extends this by illustrating the interaction between virtual workers (bots) and machinery. In this setup, both individual and coordinated behaviours are programmed to simulate real worksite dynamics, such as proximity interactions, avoidance, or simultaneous tasks. These interactions are configured through Unity’s Animator and NavMesh systems, allowing the simulation of realistic workflows and possible risk scenarios.

Together, these figures demonstrate how immersive environments can be used not only for visualisation, but also for behavioural simulation and risk prediction, providing a testing ground for proactive safety strategies and collaborative task planning in digitally modelled construction sites.

Thus, it is possible to recreate the real behaviours of workers and machinery in working environments. However, an even more relevant element is interactions with other workers (a key aspect in safety management). Figure 24 shows the interaction between a main user and a bot. Conversations, behavioural changes and indications, among other dynamics, can be established to study the worker’s behaviour and educate them about safety.

On the other hand, multi-user simulations can be recreated so that different workers interact in XR environments, allowing the study and analysis of collective behaviour. This aspect is relevant since collective responsibility in safety issues is a key aspect. The human factor, and the consequences of its actions, can be formed in the virtual environment, with implications for other virtual bots or other “real players”.

4.1.13. Weather and Other Variables

Extended reality allows for similar aspects that in reality are not possible to see or represent in a controlled way. On construction sites, the weather cannot be controlled and creates uncertainty in the performance of certain tasks, which can delay planning processes or affect construction processes. Considering these variables in XR environments allows users to prepare for these scenarios by planning proactively and realistically. Figure 25 shows examples of how these variables can be addressed from a safety perspective by climate, luminosity, and the effect of natural or artificial light or the sun in both the construction and operation stages of a project. These considerations make it possible to plan the construction, for example, according to the use of sunlight, or to avoid tasks where direct sunlight can affect workers, depending on the time of year and the position of the building under construction.

4.1.14. Gamification/Storytelling

The use of extended reality has shown advantages over traditional training methods. However, as has happened on other occasions, such as with the use of multimedia content, the simple use of new devices can generate interesting experiences as technological curiosities, but they are short-lived once the initial novelty has worn off. This is why it is necessary to further explore the purposes of these experiences and to incorporate relevant elements so that the contents achieve an effective culture of prevention among workers in the construction industry. One of the important aspects within the construction industry is achieving an awareness among workers of safety issues and forming a culture of safety that aims to achieve the global goal of zero accidents.

In view of this, especially in terms of training, the concepts of gamification and serious games have become relevant. Gamification refers to the implementation of game mechanics in various situations, in this case, in serious games (training games), where video game strategies are incorporated into XR experiences to capture users’ attention and motivate them to meet objectives [81]. Various gamification elements can be implemented, such as rewards, sounds, timers and goal counters, as can user interaction strategies that challenge the user to achieve a goal, earn rewards and surpass themselves or other competitors, among other strategies [82].

Figure 26 shows an example of gamification elements for risk identification. Counters of correctly or incorrectly identified risks are displayed along with a timer to determine the time it takes to identify all the risks in the simulation. These counters (time and score) can also be used to establish rankings and generate competition between users, for example. A mechanism of interaction with the user is also shown, through ‘Pick up elements’ and ‘Dialogue boxes’, which allow communication with and questioning of users. From an XR technology perspective, there are multiple options for implementing gamification elements, depending on the designers’ strategies.

Once the advantages of extended reality have been internalised (immersion and first-person interaction in hyper-realistic scenarios, and technologically attractive elements), it is necessary to move beyond the technological enchantment to return to the purpose of the experience: the message to be transmitted. In this approach, the use of storytelling makes it possible to study the role of emotions and cultural factors as determinants of people’s behaviour. Because storytelling is familiar from cinema and theatre as a way to convey feelings to audiences, in extended reality it helps to deepen the users’ awareness of the content they are experiencing as protagonists who are responsible for the outcome of the story and thus have an emotional investment beyond the theoretical content on safety protocols.

Storytelling refers to the art of telling a story with elements that generate emotion and transport the recipient to experience and learn from what is told. Storytelling helps to develop effective and selective attention and to improve reasoning skills; it is also a privileged instrument to develop cognitive skills [41,83]. It is a powerful and proactive method for sharing lessons about unwanted events to engage organisations and workers, aiming for optimal safety learning. Storytelling is intended to include sensitivity, trust and personal content [84]. When these elements are applied to a story (written, oral, immersive or in another form), it becomes an effective delivery method that influences the behaviour of the receiver in general and modifies the behaviour of workers in risk prevention [85], particularly in the area of safety training.

Storytelling is made up of four basic elements: (1) the message, (2) the conflict, (3) the characters and (4) the plot. These generate a good story that connects rationally and emotionally with the viewers. The message must be persuasive and exciting with a focus on the teaching that needs to be conveyed. The conflict is the main component because it is intended to generate the audience’s interest and, therefore, it must be related to the activities they perform. On the other hand, the characters experience the conflict and undergo a transformation afterwards, so they must have similar attributes to those of the target audience so that the audience identifies with the message and becomes emotionally involved with the story [85]. The plot facilitates the understanding of the story through its basic elements: the beginning, the complication, the climax, the denouement and the learning process. For the narrative to be complete, engaging and compelling, it must contain relevant and generalised information, detailing a clear cause–effect relationship through events familiar to the audience. The story must generate empathy in the viewers, and they must feel empowered to control the story [86].

Following these guidelines, Figure 27 shows a proposal of how to use storytelling elements in extended reality experiences. It details the steps to follow to construct a training experience that generates safety awareness but that can also be incorporated into XR experiences for various purposes.

According to the proposed storytelling in extended reality flow,

Table 2. Stages, description and examples of extended reality + storytelling.

Stage	Description	Example (Summarised)	Example (Full)
Exposition	Describes the situation and setting, situates and describes characters, describes their roles (positions) and activities to be performed.	The narrator is bored during a risk prevention talk but is distracted by the arrival of a new coworker, Tadeo, who had just consumed cannabis, and is now asking questions about work and life, leading the narrator to suspect that he will have to teach Tadeo every activity.	It was a bright, hot day in May, the clock was already striking 9:00 a.m. and my sleepy mind could only imagine the end of the day. José, the risk preventionist, a dark-haired guy in his 40s, began his talk like a recipe book, always the same phrases, the same advice. It was useless to focus my attention on his words. I could only think of his intrusive and inquisitive behaviour that haunted me every day. As I struggled to focus my attention on the voice of the friendly José, the new boy, Tadeo, appeared next to me. His eyes and smell gave away the fact that he had just consumed cannabis, and it was inevitable to remember the times when I didn’t care about anything but having a good time. It didn’t take a second for Tadeo, whom I had never seen before, to look for me among the group of people listening to José’s recipe book. He started asking me questions about my work and my life, which was almost like an escape route for me. At that moment, I sensed that he would be part of my crew and that I would have to teach him every activity.
Conflict	Variables, events and/or barriers influencing the activities to be carried out	The narrator reluctantly agrees to supervise Tadeo, a new colleague, and takes him to put up mouldings on the fifth floor amidst noise and dust, while thinking about his daughter’s upcoming birthday.	The eternal 20 min chat had already passed, and I had a strange feeling that it was going to be a long and difficult day. Nicolás, my supervisor, calls to remind me of the previous week’s pending activities, and informs me that Tadeo would be under my supervision. There was nothing surprising in his information. As if I had any kind of choice, Nicolás asks me if I have any objections to working with Tadeo, and I quickly and automatically say no, aware that it would mean extra work for me. The temperature had already risen and sweat was pouring down my temples; I was only thinking about the end of the day, as it wasn’t the best day to work. Quickly and with a strong tone I tell Tadeo that we have to go and put up the mouldings of the walls on the fifth floor, starting from the outside. Tadeo smiles at me. There was hot air and dust coming from the lower floors. My ears were ringing from the noise of the demolition hammer, the blowers, the hustle and bustle of the workers, machinery and the traffic on the street; in my mind, I just wanted to finish the day and get to my daughter’s birthday. She was turning 9 and I was looking forward to seeing her smile as she opened the cake I had promised to buy her on the way back home.
Climax	High-impact event, action to resolve the main conflict.	The narrator and Tadeo prepare to install a moulding, but Tadeo doesn’t understand the safety procedures, causing the narrator to fall from a height of over twelve meters.	Tadeo and I set up to receive the outer face of the moulding and install it. I pointed out to him to secure his harness to the lifeline and the procedure to be followed once the moulding was in place. But, Tadeo did not understand my words. I untied my harness, as I usually did, and walked over to him to indicate what we would do. I fell from a height of more than twelve metres.
Resolution	Reflection on the event, consolidation of acquired knowledge.	The narrator is haunted by the regretful consequences of one careless second, and wonders what could have been different if he hadn’t made that mistake.	The carelessness of 1 s will haunt me for the rest of my days. And I keep asking myself what if just one event had changed, what if I hadn’t untied my harness, what if I had talked to Nicolás about Tadeo’s condition at that moment?
Learning	Closing sentence of the story.	The narrator reflects that they can offer a lesson for others to learn from: “It’s better to change a second of your life, than your life in a second”.	I may not be able to change the past, but I can leave a small lesson for those to come, “better to change a second of your life, than your life in a second”.

describes the stages to be considered within the creation of extended reality experiences based on the storytelling aspects, and includes an example where all the required narrative elements are applied. To simplify, a single sequence of decisions is described. However, when developing the extended reality experience, it is natural to work with decision trees and the successive narratives associated with a whole set of possibilities, similar to designing video games.

The deployment of narratives in XR experiences is not simple. It requires aspects associated more with workers’ psychology and behaviour, which are humanities disciplines. However, from the perspective of technology and pedagogy, gamification strategies can help to deploy storytelling, communicate with users, attract them and generate elements that lend greater significance to the experiences. This makes the ‘story’ meaningful within the learning process and helps users internalise it.

The inclusion of gamification and storytelling elements in XR-based safety training is not merely a matter of enhancing user engagement; it is based on well-established evidence from educational psychology, cognitive science, and risk communication. Research has shown that emotional engagement, narrative immersion, and interactive challenges significantly enhance memory retention, motivation, and behavioural change, key objectives in safety learning environments. In high-risk industries such as construction, where traditional training often fails to achieve deep awareness or long-term behavioural transformation, gamified learning and narrative-driven simulations offer a promising alternative. These strategies help bridge the gap between theoretical knowledge and real-world action, fostering a culture of safety that is experiential, emotionally resonant, and self-driven. Accordingly, our approach incorporates these elements not as add-ons, but as integral components of a pedagogically and behaviourally informed training methodology.

4.1.15. Construction Progress

Extended reality should be useful in the virtual planning and simulations deployed on site, ensuring their real implementation, in association with the safety monitoring and control process, which mainly correspond to the construction and operation stage. Various mechanisms can be used to capture updated data from the construction site ranging from manually, where a manager updates the models and deploys all the models and data, to Construction 4.0 technologies and automation options (full or partial) to collect, analyse and update the XR environments.

Figure 28 shows an example of manual synchronisation of the XR environment. It includes a capture of the work site information (a photograph taken by drones), its update in the BIM model (defining a phase for the planned week, according to the actual progress of the project), and the update of the XR environment according to the progress of the project, arranging all the necessary elements (environments, bots and protection elements, among others). The safety monitoring and control process requires not only re-planning in virtual environments but also deploying safety processes on the worksite. Mixed reality and augmented reality technologies make it possible to superimpose the projected virtual models on the actual progress of the project. This synchronisation process supports real-time decision-making and strengthens the connection between virtual planning and physical execution, enhancing both coordination and safety management.

Figure 29 shows the verification of the CPEs on a project under construction. This process is carried out by manual verification and supported by real–virtual comparisons. However, because image recognition techniques are integrated with augmented reality techniques, this process will be automated.

4.2. Safety Management Tasks in XR Environments

The following is a breakdown of the four groups of tasks in safety management. The models, data or processes shown in Section 3 according to the framework proposed in Figure 6 allow for building the XR environments to develop the tasks shown in the following sections. The processes to deploy these tasks are cumulative, and the information will contribute to the deployment of the multiple safety management tasks shown (or others that are of interest). Examples are shown to clarify the use and implementation of the required safety management tasks throughout the project life cycle.

4.2.1. Safety Design/Safety Review

These tasks incorporate analysis, design and review of the safety aspects of the project at an early stage according to the guidelines of Prevention through Design. In a project analysis, the aim is to identify how to develop a safer project and influence the design and construction processes. Thus, based on BIM and project risk analysis models, XR environments already allow for first-person visualisation and interaction with these environments, enabling better identification of risk areas and aspects of project design that could affect health and safety.

Figure 30 shows an analysis and walk-though of a BIM model in the virtual environment using virtual reality goggles (in this case, using Autodesk Revit, Unity 3D and HTC Vive goggles). Setting up a review walk-through directly on BIM models is one of the simplest tasks to implement because it can be limited to interactions with a 3D model with relatively few enhancements or levels of interaction (especially if it is a spatial inspection).

Figure 31 shows an example of the Unity Reflect tool performing a measurement and reviewing the information in a virtual environment with a model superimposed on reality using augmented reality (on HTC Vive goggles and an Android smartphone).

For this type of function, commercial software allows for agile development. Associated with Autodesk Revit and Unity 3D, the Unity Reflect software allows a direct link between Revit and a virtual-reality and augmented-reality viewer, with functionalities focused on the information review of virtual elements and measurement. Although its functions are limited, it is a useful tool to quickly produce a BIM model in a functional XR environment. Although it is a generic tool, it is possible to use it for safety management purposes.

However, further functionalities can be implemented to obtain more information or to exploit the potential of XR tools. This will depend on the programming resources to generate the immersive environments, together with the needs that the safety team establishes for the project, integrated with other general uses for the development and management of the project.

4.2.2. Safety Data and Model Preparation/Protection System and PEE Selection

In this task, safety-related information and management models are prepared in accordance with regulatory requirements. The aim is to be able to establish measures to avoid accidents associated with risks that could not be avoided. With this, the planning of collective protection elements (CPE) will be required, together with the planning of individual protection elements (PPE). Both will be associated with the construction processes of each stage of the project, together with respective work dynamics. Figure 32 shows an example of the implementation of this task, ranging from the use of BIM models, risk analysis (depending on the risk to be assessed), the incorporation of the CPEs and the development of the XR environment, where the CPEs will be placed or verified, through virtual walkthroughs. This workflow demonstrates how safety planning can be digitally modelled, tested, and refined before on-site execution, improving both accuracy and anticipation of hazards. This process exemplifies a comprehensive integration of BIM, risk data, and XR visualisation, supporting early design decision-making and reinforcing the proactive implementation of safety measures in complex construction environments.

Figure 33 shows an example of a PPE verification task for review and training purposes. In this, the safety officer can review the virtual worksite and verify the PPEs of each of the workers in the different work areas and in the different tasks they perform. In this way, it can be verified that for each task all workers have their corresponding equipment. These tasks can be used to review and assign PPEs to each task, to plan their use at each stage of the project, and to train safety officers.

4.2.3. Safety Planning and Scheduling Integration/Safety Planning Review and Safety Training

It is important that the safety planning is coordinated with the work schedule and schedule updates, especially if unplanned parallel tasks are performed. Thus, the integration of schedule and project progress can be incorporated by adding the XR layer for safety coordinated with the project’s progress management (for overall project management). For example, Figure 34 shows an example of the different stages of a project. For each stage, the BIM model has been updated according to the tasks for a specific period. With this type of BIM model, the updated XR environment can be adapted to the current state of the project and the worksite can be reconfigured accordingly, in addition to the required bots, machinery and dynamics.

In this way, safety management actions will be carried out in updated XR models and environments, and with this, the revision tasks of the protection elements will be specific to the stage of the project. Along with this, the training and planning of risky tasks will be performed in a “safe” virtual environment, with the same characteristics of the real site, at the same construction stage and with the same recreated contexts. This is an important aspect as most XR experiences are static or impersonal, i.e., they are designed for fixed or standard stages of the construction process and in settings other than the specific work site of a particular project. Therefore, replicating the conditions of the specific project gives a higher degree of significance, for the management itself and even more so for recurrent training.

In line with this aspect, Figure 35 shows an example of training at the specific stage shown in Figure 34. It shows a user walking around the job site, interacting with other workers, subject to the specific contexts of construction logistics, project height (e.g., for vertigo assessment purposes) and other aspects of interest. From the worker’s point of view, the training allows them to assess their performance, fears and capabilities. However, as Figure 35 shows, a safety officer could look at the worker’s behaviour from different points of view and analyse the causes of the worker’s behaviour. Figure 36 shows an example of a worker’s decision-making in this same context, and from a ‘bird’s eye view’ the safety officer or the worker can assess their behaviour, maximising the tools to support their decision-making and encourage safe behaviour in the workplace.

While extended reality is recognised for its ability to replicate the reality and temporality of projects, a more relevant aspect is the ability to see or do what is not possible in reality. For example, Figure 37 shows a virtual environment that aims to raise awareness of the physical distance workers need to maintain for COVID-19 prevention purposes. This cannot be seen in reality. However, it can be seen in a virtual environment, and in addition, approach alerts can be generated. This is a simple example, but it can be extended to multiple information displays in any of the XR technologies.

4.2.4. Operation and Construction/Daily Inspections and Records

Linked to updating and monitoring the overall project planning, the XR environment will monitor the correct implementation of safety measures on the worksite, update models, manage documentation and review the actual environment. For example, Figure 38 shows the verification of collective protection elements using augmented reality. According to the BIM model of the project, the stage and the work zone, the CPEs are superimposed in reality using AR to verify their presence or absence to ensure correct installation and to react in time to the absence of any CPE. The information can be collected, synchronised with the BIM environment and managed to activate the company’s protocols for a prompt response by the safety officers and managers.

4.3. Validation Through Implementation Examples

The validation of the proposed XR framework was conducted through a series of illustrative implementation examples integrated throughout this section. These examples represent practical scenarios that reflect typical safety management tasks across the construction lifecycle, allowing the framework to be tested in real or simulated contexts. Each example demonstrates the applicability of specific components of the framework—ranging from BIM model integration, safety data structuring, and planning coordination, to immersive walkthroughs, safety training, and real-time inspection using AR. The examples were selected to cover different project stages (design, planning, execution, and monitoring), various XR technologies (VR, AR, and MR), and a range of safety-related activities (e.g., hazard identification, PPE verification, and site inspection).

By applying the framework to these diverse cases, its internal coherence, logical structure, and adaptability to different project types and digital maturity levels were assessed. The implementation examples also served to verify the technical feasibility of integrating BIM, safety data, and XR interaction, while illustrating how XR environments can enhance safety awareness, user engagement, and proactive decision-making. Although this validation does not replace large-scale empirical testing, it provides a strong conceptual and functional validation of the framework. The practical insights obtained from these examples support its relevance and replicability in real-world construction settings, laying the foundation for future pilot projects and standardisation efforts.

5. Discussion

5.1. General Discussions and Practical Applications

From a theoretical perspective, the proposed framework contributes to bridging a critical gap in the literature by offering a structured and integrative approach to XR deployment for safety management. While most existing studies present isolated use cases or conceptual models, this work systematizes findings from previous research and aligns them with methodological principles drawn from Construction 4.0, human–machine interactions, and digital transformation. It also provides a new lens for understanding safety as a multidimensional and dynamic process that evolves throughout the lifecycle of construction projects. On the practical side, the framework offers a replicable and adaptable strategy that practitioners can implement in diverse project settings. Its modular design facilitates integration into existing digital workflows (e.g., BIM, Lean, ISO 45001 compliance), while the illustrative examples serve as a bridge between academic insight and field applications. The framework supports decision-making, risk anticipation, and stakeholder engagement through immersive visualization and simulation capabilities—functions increasingly demanded in modern construction environments.

The proposed framework for an extended reality solution for holistic safety management in construction represents an alternative to contribute to the reduction of accidents on the construction site with a view to achieving zero accidents. Although an XR deployment such as the one shown integrates different disciplines, uses, stages and efforts within the projects, this is a reliable alternative, even more so given the digitalisation trends in the sector. It is relevant to discuss the practical applicability of the proposal and its potential impact.

• Implementation efforts. The development and implementation efforts of an XR environment for safety management appear to be large, but at first glance, the efforts and costs may be significantly higher than the benefits obtained (depending on the perspective one has). However, although the focus of this research is on safety management, XR developments have multiple uses in the AECO industry. Extended reality is one of the relevant technologies within Construction 4.0 (and even more so, within the context of Industry 5.0). Therefore, its deployment in the sector is only a matter of time. In that sense, development efforts that appear to be safety-specific will be used generally in construction project management, and the safety aspect will be added to those other uses. The generation of environments, the update and synchronisation of progress and the different tasks and construction processes, among other elements, will already come from these other areas of management. Therefore, development efforts in safety management will occur in these other areas.
• XR trends and ‘virtual universes’. Current contexts, but more importantly, future trends, point towards the use of XR technologies in all industries, and in people’s everyday life. It is relevant to consider that the different virtual universes (metaverses) and XR technologies are today a global trend. Although it seems that their widespread use has not been as rapid as expected, nobody rejects the idea that it will be the way in which humans will interact with virtual environments as well as real ones. The framework proposal therefore represents a relevant effort to prepare the construction sector to join these new forms of interaction, with a view towards the development of a ‘Construction metaverse’.
• Safety vision. Regardless of the effort involved in XR implementation or the resistance to change that characterises the AECO industry, the challenge of reducing accident rates towards zero accidents requires the implementation of new methods and technologies that can effectively address this unresolved problem in the industry. As long as human workers continue to interact in the workplace, especially in performing complex and risky tasks, prevention measures and their proper management and training must be rigorous, and XR technologies help to achieve this. However, according to the trends of incorporating robotisation and new technologies in the workplace, new strategies will be necessary to manage safety in these new contexts where human and robot workers coexist, and where new risk scenarios may occur. A clear example of this is the use of drones, where noise or the simple presence of a drone can reduce the worker’s concentration and generate potential accidents. But if we look further into the future, where only robots and automated equipment interact in the workplace, without the physical presence of human workers, the implementation of extended reality today represents a preparation for these scenarios in the sector, and in turn, an opportunity to gain experience in safely interacting with the automated workplace.
• Adaptability to project types, scales, and regional contexts. While the framework was tested and illustrated with examples of medium-scale building construction, its structure and logic were designed to be adaptable to different types of projects (e.g., infrastructure, industrial facilities), various scales (small to large), and diverse regulatory contexts. The modular organisation of model packages and data inputs allows for a flexible configuration according to project-specific constraints, regulatory requirements, or levels of digital maturity. For example, in projects operating under stricter safety codes (such as those aligned with ISO 45001 or national equivalents), specific rule-based datasets and compliance checks can be integrated into the XR environment. Likewise, the framework can be scaled down for smaller projects using simplified BIM inputs and lightweight XR implementation or expanded in large-scale infrastructure using cloud-based coordination and sensor integration for real-time data flows. Future work should explore these adaptations in more detail, particularly in projects located in regions with varied safety legislation or technological infrastructure.
• Risks and implications of XR technology failure in safety-critical contexts. While extended reality technologies offer significant advantages for improving safety awareness, training, and planning, their implementation in safety-critical contexts also involves inherent risks that must be acknowledged. A failure in XR applications—whether due to technical malfunction, low simulation fidelity, misalignment with real site conditions, or user misunderstanding, can lead to incorrect decision-making, false confidence, or a misjudgement of risks by workers and managers. For instance, if a virtual training scenario omits critical hazards or does not replicate realistic constraints, workers may adopt unsafe behaviours on the job site based on incomplete or misleading information. Additionally, over-reliance on XR tools without adequate validation may weaken traditional safety protocols or reduce vigilance during real-world operations. Latency, software bugs, inaccurate data integration (e.g., outdated BIM models), or inadequate user feedback mechanisms could compromise the reliability of XR systems. Therefore, it is essential that XR implementations for safety are rigorously tested, validated against real conditions, and continuously updated. XR should be considered as a complement, not a replacement, for established safety practices. Future research should focus on developing risk assessment models specifically for XR systems and incorporating fail-safes, redundancies, and cross-validation mechanisms within immersive training and planning environments.

The proposed framework stands out from existing literature by moving beyond conceptual or isolated XR applications. It delivers a technically detailed and operational structure that supports holistic and lifecycle-wide safety management. Unlike many previous works limited to VR-based training modules, our approach provides a replicable and theoretically grounded methodology that links BIM, Lean Construction, and digital twins with immersive XR experiences. This positions the framework as a practical tool for digital transformation, while also contributing to academic discussions on technology adoption, human-centred safety design, and the integration of emerging technologies in complex socio-technical systems. The discussion and examples presented demonstrate not only the practical feasibility of the framework but also its potential to inform standardisation efforts, support policy formulation, and guide future research into XR validation and implementation in the AECO industry. Thus, the proposal serves as a scalable and transferable solution, aligned with the broader transformation towards Construction 4.0 and beyond.

5.2. Limitations

The proposed framework represents an effort to demonstrate extended reality’s potential in safety management, and through the implementation examples, the proposal’s reliability has been demonstrated. However, in light of the research objectives and the results shown, the limitations of this work must be considered.

• While various uses and tools are shown, the developments have been focused on specific technologies, software and equipment. Although the focus of the research is to describe a technologically neutral framework, it is necessary to deploy its functionalities in diverse technologies and tools to identify limitations or greater potentials through other unidentified tools (or new technologies or future equipment).
• The proposed uses for the technologies throughout the project life cycle are framed as part of a holistic vision for construction safety management. However, the proposal has not been deployed in a complete project, which would guarantee its effectiveness and measure its results. Therefore, being able to generate such an ideal scenario, or in principle, to simulate or partially incorporate the uses in a real, entire project, would represent an interesting opportunity to evaluate the integration and effectiveness of the proposal.
• In general, massive tests of the implemented functionalities are required to analyse the effectiveness of the proposed tools in the field, to measure and analyse their impact on safety management and to establish indicators to monitor the relationship with accident reduction.
• This study does not include quantitative validation through safety performance indicators, which remains a limitation. Future work should incorporate measurable KPIs to assess the impact of the framework in operational settings.

6. Conclusions and Future Directions

Considering the construction safety issues and trends in the construction sector, a framework for an extended reality solution for holistic safety management in construction has been proposed. It responds to the trend of using extended reality in the AECO sector and seeks to fill the unresolved gaps associated with standardisation in the development of XR experiences for safety, the entire project life cycle and the integration of the different uses and factors that affect safety. Thus, the framework considers a holistic vision of construction safety management considering best practices in safety management; an approach from Construction 4.0 and methodological trends in the sector, incorporating BIM and Lean Construction, with Construction 4.0 technologies; and extended reality as a paradigm of human–machine interactions in new contexts in virtual work environments.

In direct response to the research objectives, the proposed framework integrates a holistic and lifecycle-based approach to safety management, supported by BIM, Lean Construction, and PtD principles, and is deployed within extended reality environments. It operationalizes this vision through a modular and adaptable structure that allows the integration of safety-related data, models, and XR tools across different project stages. The implementation examples illustrate how XR environments can be used not only for training, but also for planning, inspection, real-time decision-making, and monitoring. These contributions move beyond isolated applications, offering a replicable and scalable methodology that supports both current needs and future developments in Construction 4.0.

The proposal incorporates a broad view of the project life cycle from a safety perspective. The inclusion of safety aspects from the beginning of the project (Prevention through Design) is facilitated by XR environments and the incorporation of BIM. The integration of BIM environments to XR developments allows synchronization with project management, aligned to Lean planning best practices. In this way, XR environments are no longer considered static models or generic construction simulations but are transformed into updated platforms for specific projects. Safety management is performed according to the progress and particularities of each project, incorporating the necessary variables to replicate all conditions and expanding the management potential beyond what reality allows.

This proposal allows professionals in charge of construction safety to learn how to integrate XR tools into their traditional workflows and migrate to these new environments according to their needs and work stages, always with an eye on the broad vision of safety management throughout the entire life cycle. In this way, the explicit link to the trends of Construction 4.0 also shows the integration sought in the technological and methodological deployment. Therefore, this is not an isolated proposal, but one fully integrated into the way the sector is working and will work in the coming decades.

The efforts to deploy XR are aligned to the current and future trends of working in virtual environments. The future development of a ‘Construction metaverse’ will require the integration of various tools, and skilled professionals and good management practices will be equally important in these new work environments. Although the framework proposal and developments shown do not directly consider virtual work environments (such as the metaverse or NVIDIA Omniverse, for example), they represent an effort to adapt the sector to these new trends.

A relevant aspect of the framework is its capacity to adapt to different project types, scales, and regulatory contexts. Its modular design and flexible integration of data flows allow it to be customised for small, medium, or large projects, and to align with varying degrees of digital maturity. Moreover, its structure facilitates compliance with local or international safety regulations, offering a foundation for future standardisation. This adaptability strengthens the framework’s generalisability and positions it as a practical tool that can be adopted and refined in diverse construction scenarios. Future applications should focus on validating these capacities in a variety of real-world environments.

According to the limitations of the work, future lines of research should focus on the field implementation and validation of the proposed framework in real construction projects, across different types, scales, and regulatory contexts. This includes the extensive testing of its functionalities under operational conditions, the assessment of user experience and adoption by safety professionals, and the refinement of the XR components based on practical feedback. Moreover, to strengthen the justification and reliability of the approach, future work should include independent peer review and external evaluation by experts in construction safety, XR development, and digital transformation, allowing for a broader validation of the framework’s methodological soundness, adaptability, and impact. This step will be essential to move from a tested prototype to a standardized, generalizable solution that can be adopted at scale in the AECO sector. Future research should also address the evaluation of ergonomics, cognitive load, and user experience, particularly considering construction workers with varying levels of digital literacy. Incorporating these dimensions through user-centred design and testing will be essential to ensure that XR solutions are inclusive, accessible, and effectively adopted in real-world construction contexts. Future studies should focus on the development and application of quantitative safety indicators to evaluate the real-world effectiveness of the proposed XR framework, such as accident reduction rates or hazard detection performance.