AI Act High-Risk Robotics Classification: Compliance for Operators

The EU AI Act pursues a risk-based approach. Robotics systems frequently fall into the high-risk category when they are deployed in sensitive areas or constitute safety components. According to Annex III of the regulation, this includes applications in critical infrastructure, healthcare, and law enforcement. An autonomous transport robot in a hospital or a surveillance robot on a factory site often meets these criteria directly. The classification means that the system must meet strict requirements before being placed on the market and throughout its entire lifecycle. This includes a comprehensive risk management system, detailed technical documentation, and the guarantee of human oversight. Operators must understand that responsibility does not lie with the manufacturer alone. As soon as a robot is integrated into a specific workflow, conformity must be demonstrated for precisely this use case. werob uses the Spec Engine to take these requirements into account as early as the planning phase. Within 48 hours, a workflow is translated into a specification that maps both the operational goals and the regulatory guardrails of the AI Act. This prevents expensive misinvestments in hardware that later has to be shut down due to a lack of compliance.

A decisive factor for classification as high-risk AI is the link to the new EU Machinery Regulation 2023/1230. This becomes binding on 20 January 2027 and replaces the old Machinery Directive. The AI Act stipulates that AI systems that take on safety functions in products subject to third-party assessment are automatically classified as high-risk. Because the Machinery Regulation prescribes such an assessment for many autonomous robots, a double compliance obligation arises. For operators this means: a robot without a valid conformity assessment procedure under the new regulation may no longer be newly put into operation from January 2027. werob acts here as the local systems integrator that paves the compliance path for more than 44 OEM partners. Especially with Asian manufacturers, there is often a gap between technical performance and European regulation. werob closes this gap through the live Cockpit, which monitors the regulatory status of the entire fleet in real time. The 4-dimensional traffic-light system in the Cockpit immediately indicates when certifications expire or when new regulatory requirements necessitate an adjustment of the software specification. This is essential for companies that want to operate scalable fleets in several European countries.

In residential care, the regulatory density is particularly high. In addition to the AI Act, systems must meet ISO 13482 for personal care robots. The German care home inspectorate also imposes specific requirements on the use of autonomous systems in resident areas. A violation of these conditions can lead to the withdrawal of the operating license. werob has already successfully implemented projects such as that at Korian Germany, where double-digit cost relief was achieved in the first year. A concrete example is the medication round: by automating this process, cost relief of €92,000 per site per year is realized. The AI classification plays a role here when the robot makes decisions about the allocation or monitoring of patients. werob ensures that the deployed robots, such as models from Keenon or Pudu, have the necessary connectors into systems such as PointClickCare or MatrixCare. This integration enables complete documentation, which is mandatory for compliance under the AI Act. The operator thus receives not only functioning hardware but a legally compliant overall system that withstands the strict examinations of the supervisory authorities.

In the area of security and logistics, additional regulations apply, such as the German security guarding ordinance (BewachVO) and IEC 62443 for industrial cybersecurity. An autonomous robot carrying out a yard patrol generates cost relief of about €68,000 per year. If this robot, however, uses AI for facial recognition or behavioral analysis, it falls under the strictest categories of the AI Act. In some cases, real-time remote biometric identification may even be prohibited unless narrow legal exceptions apply. werob supports operators in choosing the robot specification so that it maximizes operational benefit without straying into regulatory prohibition zones. Through the Supplier Match ranking, the models that meet the necessary cybersecurity standards are selected from more than 280 rankable robots. Integration into security platforms such as Genetec takes place via pre-built connectors, which shortens the implementation time to eight weeks. The Cockpit permanently monitors the integrity of the data streams in order to meet the requirements of the NIS2 Directive and the AI Act regarding the robustness and cybersecurity of high-risk systems.

The legal risks of using high-risk AI are considerable. In the event of malfunctions or accidents, the question of liability arises between manufacturer, integrator, and operator. The AI Act provides for sensitive fines that can reach a percentage of global annual turnover. werob addresses this risk through a clear commercial model: outcome-only. Operators only pay once the robot is running in live operation and meeting the defined performance values as well as compliance requirements. This model fundamentally distinguishes werob from classic consulting firms that merely create concepts. werob assumes the technical and regulatory responsibility for the integration into the operator stack, be it SAP EWM in logistics or Opera PMS in hospitality. Through hardware agnosticism, werob is not tied to a single manufacturer and can flexibly switch to alternative OEMs from the catalog of 44+ partners in the event of regulatory changes. This offers the operator long-term investment security in a rapidly changing legal environment. Compliance is thus not a one-off project but a permanent service controlled via the Cockpit.

Implementing a law-compliant robotics solution follows a strict process at werob. While the industry usually needs three to six months for the discovery phase, werob delivers the deployable specification in 48 hours. This speed advantage is based on the Spec Engine, which was trained on data from more than 35,000 projects. After the specification, the quote follows within five days and productive deployment on site within eight weeks. This process includes the examination of all relevant standards, from the GDPR through the HACCP requirements in gastronomy to the AI Act. In hospitality, for example, the use of room service robots leads to cost relief of €112,000 per year. Here werob ensures that the AI-powered navigation in public areas meets the requirements for the safety of passersby. The connectors to systems such as Mews or Toast ensure that the robot is seamlessly integrated into the existing workflows without manual effort for data maintenance. The result is a fully automated, legally compliant solution that delivers a measurable ROI from the very first day.

A core aspect of the AI Act for high-risk systems is data governance. Operators must be able to demonstrate which data the AI was trained with and how the quality of this data is ensured. In practice this means complete logging of all system decisions. The werob Cockpit takes over this task in an automated manner. It records all relevant telemetry data and AI decisions in order to be able to present a complete audit trail in the event of an inspection by authorities. This is especially important in industries such as senior living, where the documentation obligation toward the care home inspectorate is strict. The transparency obligations of the AI Act also require that users be informed that they are interacting with an AI. werob integrates these requirements directly into the robot's user interface and the accompanying workflows. Through the connection to the operator stack, this information is delivered consistently across all touchpoints. This reduces the administrative effort for the operator and ensures that all regulatory requirements are met without manual intervention.

By 2028, werob plans to operate 2,000 robots across Europe. This growth is only possible because the system is designed for scalability and regulatory resilience. The AI Act and the Machinery Regulation 2023/1230 are not obstacles but quality standards that professionalize the market. Companies investing in robotics today must choose partners who not only know these standards but can technically implement them. With its sites in 11 European countries, including DE, AT, CH, FR, and UK, werob offers the necessary regional presence to take local regulatory specifics into account. Whether it is mowing greens on the golf course with relief of €31,000 or cleaning kitchen floors in gastronomy for €44,000: the technological base remains the werob platform. It translates complex legal texts into binary logic and operational excellence. Operators can thus concentrate on their core business while werob secures the technological and regulatory layer. The path to the autonomous organization leads through a clear compliance strategy that begins today.