By DCW ConnectAtkinsRéalis, a global leader in digital innovation and engineering, developed Virtual Site Access (VSA) to transform remote site monitoring, reducing costs and inefficiencies. Building on this success, Integrated Remote Monitoring (IRM) leverages automation, IoT, and AI to enable real-time, high-quality data collection. This next evolution ensures smarter decision-making and scalable digital solutions for complex infrastructure projects.
Can you relate to needing visual confirmation of activities or the progress of a project site, or wanting to understand its current condition? Traditionally, this requires a site visit; a time-consuming, costly, and potentially risky process. Even when visual data is captured, it is often stored incorrectly, leading to inconsistencies and outdated information.
To address this challenge, AtkinsRéalis developed Virtual Site Access (VSA) several years ago. With the successful adoption of advanced technology globally, the focus has now shifted to what comes next—autonomously collected data, available in near real-time through a cloud-based system that can be accessed from anywhere. This vision forms the foundation of VSA and drives its next evolution: Integrated Remote Monitoring (IRM).
VSA has already demonstrated a noticeable impact. On a 12-month period, it has saved an estimated 214,412kg of carbon emissions, eliminated 14,544 hours of travel time, and provided cost savings of £872,640. These figures reflect the real-world benefits of reducing unnecessary site visits while improving operational efficiency. Experts can now access critical site information remotely, ensuring data is available when needed without delays.
VSA is designed around a cloud-based architecture that enables secure data sharing, processing, and storage. It allows teams to interact with spatial data, imagery integrating seamlessly with existing digital workflows. The system ensures that data remains structured, version-controlled, and securely stored. However, while VSA enables remote access, it does not address the challenge of ensuring that collected data is structured, high-quality, and consistent over time—an essential requirement for AI-driven analysis.
Global trends see the adoption of machine learning models. Those involved in these developments strongly advocate the models are only as good as the data they are trained on. Poorly structured, inconsistent, or noisy datasets can significantly reduce model accuracy, leading to unreliable insights. This is the well-known “garbage in, garbage out” problem. To train robust AI models, data collection must be continuous, structured, and scalable. This is precisely where IRM comes in.
Integrated Remote Management builds on VSA by introducing an automated framework for large-scale data collection using robotics, IoT sensors, and remote-controlled systems. The goal of IRM is to ensure data consistency and scalability while reducing variability between datasets. By automating data collection, IRM allows for structured, high-frequency, and high-volume data acquisition, feeding machine learning models with clean and reliable inputs.
Robotics and IoT sensors deployed through IRM collect a diverse range of data types, including multi-spectral imagery, LiDAR scans, temperature and humidity readings, and live video feeds. These datasets are transmitted to cloud storage solutions designed for high-throughput ingestion and structured storage. Edge computing is used to perform initial processing steps ensuring only relevant information is sent to cloud platforms.
Once the data reaches the cloud, it undergoes further pre-processing, making it suitable for AI and machine learning applications. Cloud-based models can then use this structured dataset for predictive maintenance, anomaly detection, and automated feature recognition.
A key feature of IRM is its ability to maintain data provenance and traceability. With automation capturing vast amounts of data, maintaining metadata such as timestamps, sensor specifications, and geospatial coordinates is critical. This ensures that AI models can correctly interpret environmental variations and contextual information rather than being misled by inconsistent datasets.
Beyond AI applications, IRM enhances collaboration across teams by making high-quality data readily accessible. Since all data is processed and stored in the cloud, remote teams can interact with site data in near real-time, facilitating quicker decision-making and wider expert involvement. The ability to integrate cloud-hosted datasets with visualisation tools also allows for interactive site inspections, 3D reconstructions, and geospatial analytics, enabling engineers and analysts to assess conditions without requiring physical presence on-site.
The future of IRM lies in its scalability and adaptability. As more construction sites are equipped with automated data collection systems, machine learning models will be able to leverage increasingly large and diverse datasets, improving their predictive capabilities. Future developments may include federated learning, where AI models are trained across multiple decentralised data sources while maintaining data privacy. Additionally, advancements in sensor fusion will allow for deeper insights by combining multiple data streams into a single analytical pipeline, further enhancing situational awareness.
The transition from VSA to IRM highlights how cloud computing, automation, and AI-driven analytics are converging to transform remote site management. While VSA enables secure access to site data, IRM ensures that the data itself is collected and structured to support advanced AI applications. By leveraging cloud technology and AI-driven insights, organisations can transition from reactive decision-making to proactive, data-driven strategies.
The next time you need a critical piece of site data, imagine a world where it’s already available—collected autonomously, processed intelligently, and ready to drive insights. With IRM, that future is already taking shape.
About the Authors
Dr. Parick Geragersian is a Technology Solutions Engineer and Technical Product Owner at AtkinsRéalis, specializing in digital twinning, IoT, and AI-driven remote sensing solutions. With a PhD in aerospace engineering and autonomous navigation, their research focused on applying machine learning for UAV navigation in urban environments. They have experience developing scalable cloud architectures, integrating real-time sensor data with digital platforms, and leading technical aspects of industry projects, including nuclear decommissioning and asset management. Passionate about bridging physical and digital technologies, they are transitioning into product management to drive innovative solutions at scale.
Stephanie Boffey Rawlings is the Growth Lead for Digital and Technology solutions in the Nuclear business of AtkinsRéalis; which is where Virtual Site Access was developed. Steph has a background in immersive technologies and various CAD modelling packages that she’s utilised on multiple infrastructure projects. This has driven her to embrace and promote the various tools available in the industry to improve data acquisition and information management.