As Industry 5.0 transitions from a strategic vision into day-to-day industrial practice, the success of Extended Reality (XR) and Artificial Intelligence (AI) is no longer defined by technological sophistication alone. Instead, it depends on demonstrable, measurable value in real working conditions on the shop floor, in maintenance environments, and within safety-critical operational contexts. XR and AI must prove that they genuinely support workers by improving efficiency, accuracy, confidence, and safety, rather than adding complexity or cognitive burden.
The XR5.0 project directly addresses this challenge through the development and evaluation of six early prototype pilots, each deployed and assessed in realistic industrial environments across different domains, including manufacturing, maintenance, infrastructure, and training. These pilots were not laboratory demonstrations, but hands-on evaluations involving real users, real tasks, and real constraints, providing an authentic picture of how human-centric XR solutions perform in practice.
Rather than benchmarking isolated components or algorithmic performance in abstraction, the evaluations adopted a socio-technical perspective. They examined how workers experienced XR systems in their daily activities, how intuitively AI-supported guidance integrated into established workflows, and how factors such as usability, explainability, responsiveness, and ergonomics influenced trust and acceptance. At the same time, the evaluations identified concrete limitations and improvement needs, offering clear guidance on what must be refined before these solutions can move from promising early prototypes to scalable, deployable systems aligned with the principles of Industry 5.0.
Pilot #1: Early Validation of Human-Centric XR Design
Pilot #1 evaluated XR and AI support for human-centric product design and training in human–robot collaboration contexts. The early prototype demonstrated strong potential in terms of visual clarity, situational awareness, and perceived safety. Participants appreciated high-quality XR visualizations and real-time data integration, which supported understanding of complex systems.
However, the evaluation also highlighted limitations typical of early-stage prototypes. AI responsiveness, and onboarding effort were identified as critical areas for improvement. The findings underline that while XR training is well suited for preparatory learning, performance optimization and usability refinements are essential for operational deployment.
Pilot #2: Comparing Maintenance Support Modalities
Pilot #2 focused on the evaluation of human-centred remote maintenance and asset management, comparing three support scenarios for the same real maintenance task. The results clearly showed that mobile and voice-based solutions were currently perceived as the most efficient and intuitive for simple tasks. Remote expert support significantly increased confidence for untrained technicians.
The XR workflow scenario demonstrated clear potential, particularly for more complex procedures, but also revealed higher cognitive demand and usability challenges. Importantly, explainable AI elements improved user understanding and trust, confirming that transparency is a decisive factor in AI-supported maintenance workflows.
Pilot #3: XR and AI for Smart Water Infrastructure
In Pilot #3, operators evaluated AR-based support for smart water pipe inspection and maintenance. The early prototype combined AI-based anomaly detection, AR visualization, and voice interaction. From a technical perspective, the system showed strong rendering performance and stable interaction, even in outdoor environments.
User feedback confirmed that XR visualization enhanced situational awareness and supported decision-making. At the same time, the evaluation revealed weaknesses in AI explainability and marker precision, which negatively affected trust. These findings emphasize that AI accuracy and clarity of visualization are prerequisites for XR adoption in infrastructure maintenance.
Pilot #4: Aircraft Maintenance Training in XR
Pilot #4 evaluated an XR-based training environment for aircraft maintenance technicians, focusing on a critical component of the Wing Anti-Ice system. Participants, including junior and senior technicians, reported positive experiences regarding visual quality, interaction, and content accuracy.
Although the lack of real equipment limited objective KPI measurement, qualitative feedback confirmed that XR training improved comprehension and reduced perceived stress. The evaluation also identified usability issues related to panel positioning and interaction, particularly for users with limited XR experience, reinforcing the importance of ergonomic and intuitive interface design.
Pilot #5: Adaptive XR Training and Repair for Edge Devices
Pilot #5 assessed adaptive XR training and repair support for industrial edge devices. The early prototype integrated immersive XR applications, biometric monitoring, and Human Digital Twins. The evaluation confirmed high system stability, good responsiveness, and reliable data integration.
Participants recognized clear benefits in task efficiency, error reduction, and safety. Adaptive content personalization based on expertise and physiological state was perceived as especially valuable. While large-scale KPI tracking was not yet possible, qualitative evidence strongly suggested improved assembly and repair performance, validating the technical and human-centric foundations of the solution.
Pilot #6: Predictive Maintenance and Troubleshooting
Pilot #6 evaluated XR-supported predictive maintenance and troubleshooting using voice-controlled interaction and AI-generated guidance. The system performed reliably even under challenging conditions, such as background noise and heavy network usage, with fast reaction times that users did not perceive as disruptive.
Participants reported high trust, clear task support, and reduced cognitive load, particularly appreciating hands-free interaction. At the same time, the evaluation highlighted areas for improvement, including voice recognition robustness and richer visual information. Overall, the pilot confirmed the feasibility of XR-based troubleshooting and strong user interest in continued development.
Cross-Pilot Insights and Readiness for Industry 5.0
A holistic view of the evaluations for all six early XR5.0 prototypes reveals a consistent and coherent representation of the strengths and maturity gaps pertaining to human-centred XR solutions. Across a broad spectrum of application domains, including product design, training, maintenance, troubleshooting and infrastructure operations, participants consistently reported distinctly perceived benefits. The aforementioned improvements included enhanced accuracy in task execution, increased confidence when performing unfamiliar procedures, more effective learning and knowledge retention, and greater process consistency when compared to conventional documentation-based or purely verbal support methodologies. In the preliminary prototype phase, augmented reality (XR)-assisted guidance exhibited its potential to mitigate uncertainty, facilitate decision-making processes, and engender a heightened sense of control among employees undertaking complex or safety-critical tasks.
It is imperative to note that the findings of this study demonstrate that human-centred design principles are not discretionary embellishments, but rather constitute indispensable prerequisites for XR systems that are specifically engineered for Industry 5.0. In the field of technology, particularly in the context of artificial intelligence and voice-activated devices, there is a growing recognition of the importance of certain features in fostering trust, efficiency and safety. These features include explainable AI, which enables users to comprehend the rationale behind the recommendations received, hands-free interaction via voice and wearable devices, which enhances convenience and accessibility and adaptive guidance, which is tailored to the user’s level of knowledge and situational context. The significance of these features has been repeatedly emphasised in various studies and discussions within the academic community.