Enhancing Industrial Processes Through Augmented Reality: A Scoping Review

Abstract

Augmented reality (AR) in industry improves training and technical assistance by overlaying digital information on real environments, facilitating the visualisation and understanding of complex processes. It also enables more effective remote collaboration, optimising problem solving and decision making in real time. This paper proposes a scoping review, using PRISMA guidelines, on the optimisation of industrial processes through the application of AR. The objectives of this study included characterising successful implementations of AR in various industrial processes, comparing different hardware, graphics engines, associated costs, and determining the percentage of optimisation achieved through AR. The databases included were Scopus, SpringerLink, IEEExplore, and MDPI. Eligibility criteria were defined as English-language articles published between 2019 and 2024 that provide significant contributions to AR applications in engineering. The Cochrane method was used to assess bias. The rigorous selection process resulted in the inclusion of 38 articles. Key findings indicate that AR reduces errors and execution times, improves efficiency and productivity, and optimises training and maintenance processes, leading to cost savings and quality improvement. Unity 3D is the most widely used graphics engine for AR applications. The main applications of AR are in maintenance, assembly, training and inspection, with maintenance being the most researched area. Challenges include the learning curve, high initial costs, and hardware limitations.

1. Introduction

The optimisation of industrial processes in today’s increasingly digitalised and globalised business world demands that companies and industries implement innovative strategies to maintain their competitiveness among manufacturers, nationally and internationally, to ensure their permanence in the global market. In this context, business competitiveness is enhanced not only through traditional cost reduction, but also by adopting an organisational approach that fosters innovation and ensures long-term competitiveness to devise new ways to improve customer satisfaction [1,2,3]. This is why continuous improvement has become a crucial objective for companies seeking to improve their efficiency, productivity and competitiveness in the market. This need arises in a context of constant technological evolution and global competition, where companies must adapt to maximise their operational performance and meet market demands; therefore, the optimisation of industrial processes is vital for its ability to improve the use of resources, reduce costs, minimise production times and increase the quality of products or services offered.

Today, the concept of Industry 4.0 represents a new paradigm for organising the manufacturing process, characterised by increasing levels of digitisation and automation capable of driving the digital transformation of industries towards so-called smart factories [4]. Indeed, implementing smart production is of paramount importance to meet the demands of rapidly growing consumers in the global market to achieve high productivity and provide high-quality products at reduced cost and time. In this context, the application of emerging technologies such as augmented reality (AR) as part of Industry 4.0 emerges as a promising technology that offers potential benefits in terms of process improvement and decision making, and has emerged as a promising tool to drive these continuous improvement efforts in optimising industrial processes [5]. Modern AR technology applies virtual information to the real world through computer technologies: the real environment and virtual objects are superimposed on the same image or space in real time.

Similarly, AR has been found to be particularly beneficial for maintenance, training, assembly and inspection and quality control activities [6]. When it comes to the latter, the value and benefits of AR technologies become more evident as the complexity of the elements and assemblies to be inspected increases and AR offers a variety of applications that can revolutionise the way industrial processes are managed and improved.

However, despite the growing interest in AR technology in industrial processes, there is a lack of comprehensive understanding of how this technology can be effectively employed to optimise specific processes in different industrial sectors [7,8]. The purpose of this systematic review is to provide a clear and updated perspective on the successful application of AR in industrial processes. The review identified that studies focusing on the optimisation of industrial processes through the use of AR remain limited.

Therefore, this analysis examines different articles using the PRISMA methodology to better understand the impact and potential of AR in the optimisation of industrial processes. This is achieved through an analytical reading of publications to identify the specific industrial processes where AR has been successfully implemented. Finally, the review determines the degree of optimisation achieved through a quantitative analysis.

A systematic review of the optimisation of industrial processes through the application of AR is based on several essential reasons. Firstly, industry is undergoing a rapid transformation driven by technological advances, and AR is positioned as a promising tool to improve efficiency and productivity in industrial processes [9,10]. However, despite the growing interest in this technology, there is still a lack of complete understanding of how it can be effectively applied to optimise specific industrial processes. Therefore, a systematic review provides a comprehensive assessment of the existing research in this field. Furthermore, given the potential economic and competitive impact of implementing this technology in industry, it is crucial to have a solid and reliable evidence base to support strategic decisions and facilitate the effective adoption of this technology; thus, this project has significant theoretical, methodological and practical value.

Theoretically, it will contribute to the advancement of knowledge by synthesising and critically analysing the existing literature on the application of AR in industry. Methodologically, the use of the PRISMA methodology will ensure a systematic and rigorous review of the existing literature on the application of AR in industry [11], setting a high benchmark for future research in this field, and in practice, the results will provide clear and applicable guidelines for the implementation of AR in industrial processes, helping companies to improve their operational efficiency within their industrial processes.

2. Methodology

The methodology proposed for the systematic literature review is based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. This research strictly follows the PRISMA guidelines, which aim to help authors of systematic reviews to generate a bias-free document of high scientific quality. In its latest version, published in 2020, PRISMA incorporates significant methodological advances in the classification, selection and summarisation of studies, ensuring a more rigorous and transparent review process [11]. In the 2020 statement, it reflects the methodological advances in the classification, selection and summarisation of the different studies.

To achieve this, relevant studies addressing the implementation of AR in the industrial domain were collected. Subsequently, a detailed quantitative analysis was carried out, focusing on the collection and evaluation of numerical data obtained from these studies. This analysis determined the percentage of optimisation achieved in industrial processes thanks to the implementation of AR. By examining key performance indicators, such as operational efficiency, error reduction and time savings, a comprehensive view of the benefits and limitations of this technology in the industrial sector was provided. This methodology not only facilitated an in-depth understanding of current impacts, but also identified potential areas for future research and practical applications.

2.1. Research Questions

Four research questions were posed with the purpose of understanding the importance of implementing AR technology in different industrial processes. Through these questions, we aim to provide a comprehensive and informed understanding of the potential and limitations of AR in the industrial sector, thus supporting informed decisions on its adoption and optimisation in various applications. See

Table 1. Main questions and motivations.

Number	Question	Motivation
RQ1	What are the main benefits of applying AR in the optimisation of industrial processes?	Identify the advantages of applying AR in industrial processes.
RQ2	What is the most used graphics engine?	Identify market trends and benefit from a large support community to make informed decisions.
RQ3	In which industrial processes is it recommended to implement AR?	Ensure a strategic and beneficial adoption of AR in industrial processes.
RQ4	What challenges and limitations do industries face when implementing AR?	Identify the drawbacks that arise when implementing AR.

for a detailed summary of these questions and the rationale behind them.

2.2. Document Search

According to [12], it is recommended that, to achieve a maximum count, searches in systematic reviews should include a combination of databases and publishers. To ensure adequate search performance (i.e., count, precision and number needed to read), we assume that literature searches for a systematic review should be conducted through at least a combination of four databases. The combination of four databases improves both the sensitivity and precision of the search process, ensuring that relevant studies are not missed and that screening efforts are optimised. Therefore, the following sources of information were chosen: Scopus, an extensive citation and abstract database covering a wide range of global and regional scientific journals, conferences and books, with a strict selection process to ensure data quality and relevance; SpringerLink, a renowned scientific platform offering high-quality content to academic researchers, scientific institutions and companies, providing access to a vast collection of books, journal articles and conference papers; MDPI (Multidisciplinary Digital Publishing Institute), a leading open access publishing institute, known for its article processing fee-based business model, which guarantees the rapid publication of peer-reviewed research and ensures unrestricted global accessibility and dissemination of scientific results; and IEEE Xplore, a relevant source of information that provides access to the technical and scientific literature published by the Institute of Electrical and Electronics Engineers (IEEE), including journal articles, conference proceedings, technical standards and other important documents, and which is an essential reference for researchers and practitioners in the fields of engineering, computer science and related technologies.

The choice of these information sources ensures a comprehensive and high quality coverage of the relevant literature, allowing for a rigorous and well-founded systematic review of the implementation of AR in industrial processes [13,14,15]. First, a search for articles in databases from 2019 to 2024 was conducted. This time interval was chosen because the implementation of AR in industrial processes has advanced exponentially in recent years, so it was considered that selecting articles from 2019 to 2024 was an appropriate period to analyse the applications, advantages and disadvantages of AR in industry [16].

This temporal approach allows capturing the most recent and relevant developments in the field. The search for documents in the various databases was carried out by entering the following combination of specific terms: (‘augmented reality’) AND (‘industry’ OR ‘industrial process’ OR ‘manufacturing’ OR ‘production’ OR ‘Industrial’) AND (‘optimization’ OR ‘ Enhancement’ OR ‘ Improvement’ OR ‘Maximization’). These keywords are directly related to the research questions, taking into account that the focus of the systematic review corresponds to the applications of AR in industrial processes. This article search and selection process is essential to ensure that the systematic review includes relevant and high-quality studies that provide a comprehensive understanding of the implications and benefits of AR in industrial processes.

2.3. Paper Selection

For the selection of articles, we used the PRISMA methodology flowchart, which allows us to represent the information through the different phases of a systematic review. This includes the number of records identified, the articles included and excluded, as well as the reasons for exclusions. This diagram was divided into three phases: identification, selection and inclusion. In the identification phase, 437 articles were found in the four databases and publishers; in the first step of this phase, 23 duplicate articles were found of which 414 articles were left for the selection phase. In this phase, as a first step, 306 articles were eliminated according to the reading of titles and abstracts; then, as a second step of the selection phase, the 108 articles were analysed to eliminate 25 according to the eligibility criteria indicated in

Table 2. Criteria and descriptions.

Abbreviation	Criteria	Description
C1	Study Design	Articles related to applications of AR in industrial processes.
C2	Language	Only articles written in English were selected.
C3	Time Range	Articles published between 2019 and 2024 were selected.
C4	Significant Contribution	Articles that provide information on the applications of AR in engineering through implementation and obtaining tangible results.

.

In addition, as the third step of the selection phase, the remaining 83 articles were screened to eliminate 37 articles according to their relevance and perspective. Subsequently, 46 full articles were assessed for eligibility. Of these, three articles were excluded because they did not meet the appropriate study type for the purposes of the review. Then, using a ‘Cochrane risk of bias’ analysis, which refers to the possibility that the results of a study in a systematic review are distorted due to certain factors, we assessed the potential biases in the included studies. This rigorous selection process ensures the quality and reliability of the studies included in the systematic review. The seven bias domains were adapted to the review to five domains by the investigators [17]. These domains were:

• D1. Inclusion criteria bias: Inaccuracy of selection criteria will lead to bias [18]: According to the research objective, this article clearly delimits the topic of study and the type of literature review approach, thus establishing the keywords to be searched for. Criteria, such as period and search language, are established to control any potential bias. In this domain, it was decided to include articles that despite not being within the established time period are of high relevance to the research.
• D2. Data completeness bias: This refers to the completeness and accuracy of the data collected during the assessment of industrial processes and whether methods were applied to ensure the objectivity and validity of the results.
• D3. Methodological quality bias: To assess the methodological quality of the studies included in the review, considering aspects such as study design, sample size, blinding and other elements that may influence the validity of the results, specifically in the context of AR in industrial processes.
• D4. Evidence synthesis: Synthesising the results of the studies included in the review in a transparent and objective manner, considering the risk of bias in the interpretation of the results. This involves careful assessment of consistency and heterogeneity among studies, as well as consideration of possible bias in the interpretation of results. It allows the identification of all the components necessary to judge the risk of bias in reporting to improve the credibility of evidence syntheses.
• D5. Researcher experience bias: Allows the identification of whether he or she or the authors consciously or unconsciously influence the results of the study, either in the design of the study or the interpretation of the data, which may affect the validity of the results.

Using the Risk of Bias web application, designed to visualise risk of bias assessments made as part of a systematic review, a “Traffic Light Plot” graph of the domain-level judgments was produced for each individual article (see Figure 1), which provides a clear visual representation of the risk of bias. In addition, this tool allows the generation of weighted bar charts showing the distribution of risk-of-bias judgments within each domain, as shown in Figure 2, which facilitates a more detailed and accurate understanding of how risk of bias has been assessed for each aspect of the studies reviewed. These visualisations help reviewers and readers to effectively interpret and communicate the results of the risk of bias assessment, thus facilitating informed research decision-making.

At this stage of elimination, due to risk of bias, five articles were eliminated due to their high level of risk, since their results could have been distorted. However, two of them presented some minor concerns as they were outside the CE3, although it was still decided to include them in the research due to their relevance to the topic under study. After a rigorous selection of the articles, 38 were selected for the research at the inclusion phase, as shown in Figure 3, thus representing a robust and adequate set of studies for the analysis and interpretation of the results.

3. Results

Literature Review

Appendix A presents a compilation of various applications of AR in different industrial processes. Each entry includes the title of the article, a description of the industrial process in question, how AR was implemented, as well as the benefits and limitations observed. This information provides a comprehensive overview of the advantages and challenges associated with the use of AR for industrial process optimisation.

AR is applied in a wide variety of industrial processes to improve the efficiency, accuracy and safety of operations. In the field of maintenance, it facilitates the visualisation of procedures and interactive guides that help technicians perform complex tasks with greater precision. For example, in the maintenance of industrial robots and injection moulds, it provides visual instructions that guide the disassembly and assembly of components, improving accuracy and reducing operating time. In the assembly area, AR is used to superimpose visual instructions directly on the operator’s work environment. This includes text, videos and 3D animations that guide workers through the various stages of assembly, improving the accuracy and efficiency of the process. The implementation of AR in the manufacturing and inspection of parts allows for easy comparison of the CAD design with the actual parts and adjustment of the transparency of the 3D representations to avoid confusion.

In logistics and warehouse management, AR is applied to optimise order picking. Systems such as RASPICK use head-mounted devices that display item lists and guide pickers to the correct locations in the warehouse. It also facilitates the identification of storage locations and verification of required parts, improving accuracy and reducing search time. It plays a crucial role in worker training and education. In engine assembly and critical component repair, AR allows students to interact with virtual components in a physical environment, improving understanding and execution of complex tasks. Furthermore, in firefighting training, it provides a comprehensive and realistic training experience, combining the real and virtual worlds to improve operator readiness.

In the quality control process, AR is used to guide inspectors during the visual inspection of welds and other components by overlaying digital information directly on the physical parts. This allows inspectors to quickly compare welds to ideal standards and detect potential defects more efficiently. Finally, in the maintenance of infrastructure such as roads and bus fleets, it improves safety and operational efficiency. It provides real-time visualisations, interactive instruction and human–robot collaboration, reducing risk exposure and improving accuracy and speed of task execution.

Appendix B presents a compilation of AR hardware along with their approximate costs in U.S. dollars and the graphics engine used. This information provides a comparative view of the options available on the market, facilitating the evaluation of AR technologies according to their technical and economic characteristics.

The most widely used hardware for augmented reality applications in industrial processes includes smartphones and tablets, mainly due to their wide availability and affordable price range. HoloLens integrated with ISO14644-14 [19] also stands out as an advanced and specialised option, especially in industrial and professional environments, despite its higher cost. Other hardware such as Vuzix M400 and Epson MOVERIO BT-300 are also popular in specific niches due to their specialised features.

Analysis of the column diagram in the Figure 4 data reveals that HMDs (head-mounted displays) are the most widely used hardware in AR applications in the identified processes, with a total of 21 units. This is due to their ability to deliver immersive and hands-free experiences, making them ideal for a wide range of AR applications. HHDs (hand-held devices) also have a high usage rate, highlighting their popularity and versatility. Hardware such as cameras, monitors, laptops and projectors have a more specific and limited use, but are still important in certain AR contexts. In terms of graphics engines, Unity 3D is the most widely used due to its versatility and ability to support multiple SDKs, such as Vuforia, ARCore and the Mixed-Reality Toolkit (MRTK). This allows developers to create robust and functional applications for a variety of hardware. Other graphics engines and SDKs are also used, albeit to a lesser extent, for specific applications that require particular capabilities.

Appendix C presents an analysis of the impact of AR implementation in industrial processes. It compares the state before and after the adoption of AR, highlighting the percentage of optimisation achieved. This information provides a clear view of the quantitative benefits that AR can bring to efficiency and productivity in industrial environments.

The results show the significant benefits of applying AR in various industrial processes. The implementation of AR has shown a remarkable reduction in task execution time. For example, a reduction in design time by 45.22% and assembly by 25% was observed. Item collection time decreased by 9% for 10-item lists and 12.55% for 20-item lists. For inspection and documentation, inspection time was reduced by as much as 50 percent and fault documentation time by 66 percent. In addition, training and maintenance also showed significant reductions, with training time decreasing by 75 percent and maintenance time decreasing by 46 percent.

Productivity and efficiency also improved significantly with AR. Trainer productivity increased by 10 percent by reducing interruptions during work. Efficiency in both simple and complex tasks increased, with task completion time reduced by 30.31% through the use of HMD hardware. Additionally, the efficiency of disc brake system maintenance increased by 32.53%. Work quality also benefited, reducing the duration of the quality control process by 24.63% for expert operators and 49.08% for non-expert operators. Error reduction is another significant benefit of AR implementation. A 32.35% decrease in errors during spindle assembly and an 81.08% reduction in errors during assembly with AR warnings were observed. In addition, spot-weld accuracy increased by 52%, improving the accuracy of the work performed. Product quality also improved with AR, with product porosity reduced by 14.94% over previous studies, indicating an improvement in final product quality. In terms of costs, maintenance personnel costs per vehicle were reduced by 0.32%, representing a small but significant reduction in operating costs. Optimisation in inspection and repair processes also benefited from AR. The inspection time of highway structures was reduced by 78% and the repair of compressors with an AR MARMA system showed a reduction in repair time by 30%. Thus, it is demonstrated that applying AR in industrial processes provides remarkable benefits, including reduced task execution time, improved productivity and efficiency, reduced errors, improved product quality, and cost savings. These benefits are reflected in an overall optimisation of operational performance and increased efficiency in performing critical tasks in the industrial environment.

4. Discussion

Research Questions

The 38 articles selected above contain sufficient information to make a scientific judgment regarding the optimisation of industrial processes through the application of AR. The questions posed in Section 2 are answered below.

RQ1: Implementing AR in industrial processes offers a wide range of benefits that contribute to the optimisation and efficiency of operations. These benefits include error and time reduction [20,21], as AR enables error identification and avoidance, significantly reducing the error rate in tasks such as assembly [22] and inspection. In addition, task execution times are significantly reduced, as seen in the reduction in design and assembly time, and in the completion of maintenance tasks. AR also contributes to increased efficiency and productivity. This technology helps operators understand design procedures by overlaying virtual information and instructions on the real environment, facilitating the completion of complex tasks. The system allows both hands of the picker to remain free and reduces cognitive load, thus improving work efficiency and productivity. In terms of training, AR allows training to be conducted in a dedicated environment, minimising disruption to daily operations and improving overall efficiency.

In addition, it enables accurate and guided visualisation of procedures, reducing human error and accelerating the learning process. Maintenance optimisation is another key benefit [23,24]. The adaptive mode of AR enables users to recognise faults faster and with fewer movements, improving efficiency and reducing downtime. AR provides contextual and meaningful information automatically, enabling faster and more efficient maintenance of equipment. Improved work and product quality is evident with the use of AR, as accuracy in tasks such as welding and quality control is increased, reducing variability and improving quality standards. AR also reduces the need for additional communications between workers, streamlining the process of video annotation and other documentation procedures. In terms of cost reduction, early detection of errors during assembly through AR prevents major problems, significantly reducing repair and rework costs.

AR offers opportunities for continuous performance improvement, which can lead to significant savings in long-term operating costs. AR also facilitates improved communication and data management with online databases, allowing real-time access, editing and sharing. Similarly, AR provides a 3D visual representation of components and processes, which improves understanding and accuracy in the execution of tasks. Finally, AR increases safety by generating visual signals that warn workers of hazards in their environment, improving safety and reducing the risk of accidents. In addition, it provides a safe and controlled training environment, where users can practise emergency situations without real risks.

RQ2: The most widely used graphics engine for developing AR systems is Unity 3D [20,25,26]. Unity 3D is a widely used graphics engine for developing AR applications in video games, simulations and interactive 2D and 3D experiences. Developed by Unity Technologies, this engine allows developers to create visually stunning and highly interactive content efficiently. One of its main advantages is its ease of use, thanks to its intuitive interface and extensive library of resources and tools that facilitate the creation of graphics and the programming of behaviors. Unity 3D supports multiple platforms, allowing developers to launch their applications on operating systems such as Windows, macOS, iOS, Android, and video game consoles, among others.

The Unity 3D graphics engine is a complete development platform that provides the tools and environment necessary to create, design and run interactive and visual applications. A graphics engine includes capabilities for handling graphics, physics, audio, animations, artificial intelligence, and other essential components for application development. On the other hand, an SDK (software development kit) is a set of tools and libraries that allow developers to integrate specific functionalities into their applications. SDKs are not complete development platforms, but add-ons that are added to a graphics engine or other development environment to provide additional capabilities.

RQ3: AR is emerging as a valuable technology in a variety of industrial processes. Its ability to provide visual information and interactive guidance is revolutionising the way tasks are performed in various industrial processes. In Figure 5 an analysis is presented based on the research of the number of industrial processes where AR is applied, supported by the existing literature.

After interpreting the data provided, the industrial processes in which it is highly recommended to implement AR are as follows: maintenance, assembly, training/coaching and inspection. These processes have been supported by a larger number of articles, suggesting that the implementation of AR in these areas has been widely researched and considered beneficial. Other processes can also benefit from AR, but further evaluation based on specific needs and use cases is recommended.

AR is most widely used in the maintenance process due to a combination of unique advantages that this technology offers to improve efficiency, accuracy and safety in various maintenance tasks [27,28,29,30]. AR provides 3D animated instructions directly on the object or work area, making procedures easier to understand and reducing the possibility of errors. Technicians can receive remote assistance from expert engineers via HMD devices, transmitting live video of the work environment and receiving annotations and AR content superimposed on their field of view, which is especially useful in workshops where the physical presence of the expert is not necessary. In rail vehicle maintenance, AR is used to develop training content that visualises procedures in a 3D space, improving operator understanding and competence.

RQ4: Implementing AR in industrial processes comes with a number of challenges and limitations that must be considered. One of the main obstacles is the learning curve associated with the technology. Operators and designers may need time and training to become familiar with the tools and techniques required to use the system effectively [31,32]. In addition, the portable hardware used in AR requires significant computational power, which can be a technical challenge. Another important limitation is the initial cost of implementation is high, requiring significant investments in specialised hardware and software development tailored to the specific needs of each process. In addition, the narrow field of view of hardware such as HoloLens can limit the amount of information visible simultaneously, making it difficult to fully perceive fault icons in some situations.

The initial installation and configuration of infrastructure such as wireless sensor networks (WSNs) for real-time monitoring can be costly [33,34]. The creation of interactive 3D models and the implementation of AR technologies require a significant investment in time and money, which can be a barrier for some companies. The effectiveness of AR also depends on the quality and compatibility of the hardware used, which can result in additional costs for companies wishing to implement this technology. The AR system requires ongoing maintenance and upgrades to ensure optimal performance, which can demand additional resources and time. In addition, there may be resistance to change from workers and stakeholders accustomed to traditional methods, which may hinder the adoption of new technologies. Insufficient lighting and high dust can slow down and complicate initialisation and monitoring in the process. Implementing AR and motion capture systems is also costly, including hardware, software, and personnel training.

Lighting conditions can affect display stability, which can be problematic in industrial environments. In addition, some workers, especially those with little familiarity with the technology, may tend to ignore screen-based AR, suggesting that while AR is powerful, it requires time and training for users to become accustomed to its use [31,35]. While AR can improve the user experience once mastered, it may require additional time and training for workers to become familiar with the technology and learn to use it effectively, which can delay adoption and initial benefits. Reliance on advanced technology infrastructure, including HMD hardware and stable, high-speed network connections, can be a challenge in technology-constrained industrial environments.

Implementing AR systems can be costly and requires workers to adapt to new work methods, which could lead to resistance or an initial learning curve. The need for calibration and ongoing maintenance of projection systems to ensure that visual cues are accurate and correctly aligned with panels requires additional resources and time. In addition, potential ergonomic discomfort and visibility issues with HMD hardware, especially in low light conditions, may hinder prolonged use and negatively affect the user experience.

5. Conclusions

A systematic review was carried out applying the PRISMA methodology to ensure rigor and transparency in the selection of studies, comprehensively evaluating the effectiveness of AR in the optimisation of industrial processes. The findings reveal that the implementation of AR has generated significant benefits in several industrial areas. Among the main benefits are substantial reductions in task execution times, improvements in operational efficiency and productivity, as well as a notable decrease in errors during process execution, contributing to a substantial improvement in the quality of the final product and a reduction in long-term operating costs.

The industrial processes recommended for the implementation of AR include maintenance, assembly, training, and inspection, areas that have proven to benefit widely according to the researched literature. In terms of technology used, the most prevalent hardware is HoloLens HMDs, followed by HHDs smartphones and tablets. Unity 3D has emerged as the leading graphics engine for AR application development, thanks to its versatility and compatibility with multiple SDKs such as Vuforia and ARCore. Despite the benefits described above, AR implementation faces significant challenges such as the learning curve, which requires extensive training for operators, users and designers. In addition, the high upfront costs of AR acquisition and infrastructure, including advanced hardware and supporting networks, represent a considerable financial barrier for many organisations.