“Hypervision” with AI-supported security architecture instead of just cameras: What’s behind Morocco’s new surveillance network

Fighting train cancellations and delays: This is how AI is supposed to save the Moroccan rail network

The 96 billion dirham program (8.8 billion euros): Morocco is building a new operating system for its railways

Morocco's state-owned railway company ONCF is facing a historic transformation. With the 2030 FIFA World Cup in mind and passenger numbers rapidly increasing, the country is investing 96 billion dirhams in expanding its rail network over the coming years. However, a seemingly tiny item in this enormous budget is proving to be the true centerpiece of the railway's digital future: a feasibility study costing around one million dirhams is intended to lay the foundation for a nationwide, AI-powered safety architecture.

With 6,000 planned cameras – almost a third of which will be automated using artificial intelligence – Morocco is planning the leap to so-called "hypervision." This involves far more than just traditional video surveillance. It's about quickly preventing delays, protecting critical infrastructure from cyberattacks, safeguarding data privacy by deliberately avoiding facial recognition, and ultimately, determining whether Morocco remains technologically sovereign or becomes dependent on foreign providers. The following article examines why this very architectural decision will determine the economic, operational, and industrial policy success of Morocco's railway of the future.

Morocco's rail network is becoming an intelligent security platform: From camera project to system decision

At first glance, the planned AI-based surveillance architecture of the Moroccan state railway ONCF appears to be a relatively small digital project. The tendered contract has a volume of 1.08 million dirhams and is intended to develop a nationwide strategy by 2030. Compared to the ONCF's total investment program of 96 billion dirhams, this amount is almost insignificant. However, from an economic perspective, it would be wrong to conclude that the project is of little importance. The tender does not finance the complete technical implementation, but rather the preliminary conceptual work: architecture, standards, priorities, integration logic, investment requirements, and implementation roadmap. This phase, in particular, determines which suppliers will ultimately be awarded the contract, how open or closed the system will be, and what the long-term costs will be.

The core of the project is therefore not the procurement of additional cameras. ONCF is planning, rather, the transition from numerous individual surveillance solutions to a centrally controlled, data-driven security system. By 2030, around 6,000 cameras are to be deployed across the rail network. Approximately 30 percent, or roughly 1,800 units, will automatically analyze images using artificial intelligence. The remaining cameras will remain conventional sensors but will be integrated into the same overarching platform. This combination makes economic sense: not every location requires expensive computing power or sophisticated analytical capabilities. The key is to intelligently monitor particularly critical areas while simultaneously consolidating all relevant signals into a comprehensive overview.

The tender is therefore an architectural contract with considerable leverage. A well-designed system can increase security, reduce operational downtime, enable more targeted personnel deployment, and facilitate future expansions. Conversely, a poor design can lead to technological dependency, high integration costs, alarm overload, and unclear responsibilities. The low price of the study should therefore not be confused with the value of the decision. Compared to the subsequent hardware, software, network, storage, maintenance, and operating expenses, the planning effort is small. However, its impact on these expenses is very significant.

Security as economic infrastructure

Railway safety is often viewed as a cost center. This perspective is too narrow, because safety in rail transport is simultaneously a prerequisite for capacity, reliability, and trust. Unauthorized access to the tracks, an abandoned object, a fire, a dangerous crowd, or an obstruction on a high-speed line can cause significant consequential costs. These include train cancellations, delays, evacuations, diversions, equipment damage, additional staff deployments, and reputational damage. In an expanding network, not only do traffic volumes and revenue opportunities increase, but so do the number of potential disruptions and the economic consequences of each major incident.

The economic benefits of intelligent video analytics stem primarily from speed. Cameras don't automatically prevent an incident, but they can shorten the time between its occurrence, detection, assessment, and response. This timeframe is particularly critical in rail transport because a local disruption can quickly affect timetables, connections, and subsequent sections of track. Early detection of a person on the tracks or timely identification of an obstacle allows for controlled intervention before a safety issue escalates into a major operational incident. The benefits then extend beyond simply preventing damage; they also include reduced delays and more stable utilization of scarce infrastructure capacity.

In addition, there is the trust effect. Passengers evaluate a rail system not only by speed and price, but also by perceived safety, reliability, and orderliness. This is especially true for major train stations, heavily used commuter routes, and large international events. Morocco is preparing its transport networks for significantly higher loads, partly in connection with the 2030 FIFA World Cup. Modern security management can therefore indirectly strengthen the acceptance of new services. While it doesn't automatically improve demand, it reduces a major obstacle to using public transport.

At the same time, the benefits must be realistically assessed. Video analytics does not guarantee security. It shifts the focus from reactive observation to proactive pattern recognition. Whether this generates economic added value depends on whether alarms are reliably prioritized, reviewed by qualified personnel, and translated into effective operational processes. An algorithm without a clear operational workflow produces data, but not necessarily security.

Growth increases the pressure to act

The project coincides with a period of strong growth in Moroccan rail transport. ONCF carried approximately 55.1 million passengers in 2024, compared to 38.5 million in 2019. This represents an increase of around 43 percent within five years, even though this period was impacted by the pandemic and its consequences. Passenger numbers reached approximately 55.6 million in 2025. While growth has slowed from a high level, the long-term trend remains clear: rail is gaining importance as a mode of transport, and the planned expansions are intended to unlock a new level of scale.

Revenue development also demonstrates increasing economic relevance. ONCF's revenue in 2024 was 4.82 billion dirhams, approximately 11 percent higher than the previous year. Passenger transport generated 2.763 billion dirhams, eight percent more than in 2023. In 2025, total revenue exceeded five billion dirhams for the first time. These figures demonstrate that the railway is not merely a state infrastructure project, but a growing mobility and logistics service provider with increasing operational demands.

The development is particularly evident in high-speed rail services. In 2024, Al Boraq transported approximately 5.5 million passengers and generated around 780 million dirhams in revenue. By 2025, this number had risen to roughly 5.6 million passengers, while revenue reached approximately 848 million dirhams. Thus, while Al Boraq accounted for only about one-tenth of the passengers in 2024, it generated significantly more than a quarter of the passenger revenue. This indicates a higher average revenue per passenger and the strategic importance of high-quality long-distance rail services.

This results in a clear scaling effect for the security architecture. With increasing traffic, the number of contacts, station density, the complexity of transfers, and the value of uninterrupted operations all increase. A monitoring system that seems sufficient for today's network could reach its organizational and technical limits within just a few years. Therefore, planning up to 2030 must not only reflect current needs but also take into account peak loads, new stations, additional rolling stock, regional services, and the extended high-speed rail line.

The 96 billion dirham program

The AI ​​project can only be understood within the context of the larger investment cycle. ONCF plans to spend a total of 96 billion dirhams by 2030. Of this, 53 billion dirhams are earmarked for infrastructure and equipment for the approximately 430-kilometer high-speed rail line from Kenitra via Rabat and Casablanca to Marrakech. A further 29 billion dirhams are allocated for 168 new trains. The remaining 14 billion dirhams will be used for modernizing, stabilizing, and further developing the existing network, as well as preparing for regional rapid transit services.

The scale of the project illustrates why the security strategy must be defined early on. The future video and hypervision platform should not be retrofitted onto an existing infrastructure. It must be integrated into station planning, communication networks, operations centers, vehicles, power supply, maintenance processes, and cybersecurity. The later interfaces are defined, the more expensive retrofits will be. From a lifecycle perspective, it therefore makes sense to design the digital security architecture in parallel with the physical infrastructure development.

The overall program also changes ONCF's role. From a national rail operator with a relatively concentrated network, it is increasingly becoming an integrated mobility provider for high-speed, long-distance, regional, and metropolitan transport. More frequent, RER-like services are planned for Casablanca, Rabat, and Marrakech. This will increase service frequency and passenger turnover, while reducing the time available for operational decisions. Security systems will then need to monitor not only individual objects but also movement patterns across numerous locations and aggregate information in real time.

The expansion has an industrial policy dimension. Large railway programs create demand for construction services, signaling technology, rolling stock, telecommunications, software, maintenance, and training. The monitoring strategy can influence how much of this added value is generated in Morocco. If open interfaces, local integration expertise, and transferable standards are required, Moroccan companies can be involved in the operation, adaptation, and further development in the long term. Conversely, if a completely proprietary, all-in-one solution is procured, there is a risk that a significant portion of the digital value creation will remain permanently tied to foreign providers.

Hypervision as a new operating system

The term hypervision describes the integration of various monitoring and alarm systems into a single, overarching platform. In practical terms, this means that camera images, AI alerts, access control data, fire alarms, technical warnings, and potentially other operational data are no longer processed in separate applications. Instead, a unified overview is created, allowing events to be assessed, prioritized, and forwarded to the appropriate authorities.

The economic advantage lies in the reduction of organizational friction. In fragmented systems, personnel have to consolidate information from different interfaces. Responsibilities can be unclear, warnings are processed twice, or not forwarded in a timely manner. A central platform can reduce these media breaks. It enables standardized processes, uniform escalation levels, and traceable documentation. This allows regional control centers and a national security center to act in a coordinated manner without completely centralizing every decision.

Centralization, however, brings new risks. A platform that connects many systems becomes a critical node itself. Technical failures, cyberattacks, or misconfigurations can affect large parts of the network simultaneously. The system therefore requires redundancy, separate security zones, emergency operating procedures, and local response capabilities. Hypervision should not be confused with complete central dependency. A robust architecture must combine centralized oversight with decentralized resilience.

The sheer volume of information also presents a challenge. 6,000 cameras continuously generate large data streams. If all recordings were permanently stored and analyzed centrally in high resolution, the demands on bandwidth, storage, and computing power would increase considerably. A tiered architecture is usually more economically viable. Certain analyses can be performed directly at the camera or on local computers, while only relevant events, metadata, or selected video sequences are transmitted to the central system. This so-called edge processing reduces transmission costs and can shorten response times. However, it complicates maintenance and security management because more intelligent devices need to be operated in the field.

Where artificial intelligence creates benefits

The planned applications focus on specific, operationally relevant situations. These include detecting unauthorized access, monitoring crowds, identifying abandoned objects, and observing trains. On high-speed lines, animals, obstacles, or objects on the tracks can also be detected. In train stations, density, direction of movement, and congestion can be analyzed. The project thus operates in an area where computer vision can generate measurable benefits, provided the models are adapted to local conditions.

Intrusion detection is particularly relevant because railway lines cover large areas that are difficult to secure completely. A system can monitor defined zones and report movements that are unusual at a specific time or under specific conditions. The advantage over simple motion detection is that modern systems can distinguish between people, vehicles, animals, and environmental factors. Ideally, this reduces false alarms caused by shadows, rain, vegetation, or changes in light.

Crowd analysis pursues a different goal. It focuses less on individual people and more on density and flow. If an unusually large number of people gather on a platform or opposing flows emerge, the control center can react early, manage access, deploy staff, or adjust passenger information. The benefits are both safety-related and operational. Better distribution of passengers can stabilize dwell times and prevent critical situations at platform edges.

Detecting unattended objects sounds simple, but it's technically challenging. Train stations are dynamic environments with luggage, cleaning equipment, vendors, and changing visibility. A usable system must distinguish whether an object is truly unattended, how long it has been there, and whether it's blocking a relevant area. Without careful calibration, numerous false positives are likely. Therefore, economic success doesn't depend on the longest possible feature list, but rather on a few reliably functioning applications with clearly defined thresholds.

Here you will learn how your company can implement customized AI solutions quickly, securely and without high entry barriers.

A managed AI platform is your all-inclusive, worry-free solution for artificial intelligence. Instead of dealing with complex technology, expensive infrastructure, and lengthy development processes, you receive a ready-made solution tailored to your needs from a specialized partner – often within just a few days.

The key advantages at a glance:

⚡ Rapid implementation: From idea to ready-to-use application in days, not months. We deliver practical solutions that create immediate added value.

🔒 Maximum data security: Your sensitive data stays with you. We guarantee secure and compliant processing without sharing data with third parties.

💸 No financial risk: You only pay for results. High upfront investments in hardware, software, or personnel are completely eliminated.

🎯 Focus on your core business: Concentrate on what you do best. We take care of the entire technical implementation, operation, and maintenance of your AI solution.

📈 Future-proof & scalable: Your AI grows with you. We ensure continuous optimization and scalability, and flexibly adapt the models to new requirements.

More information here:

• Managed AI Platform

The underestimated price of false alarms

The key business metric for an AI monitoring system is not its theoretical detection rate in the lab. What matters is the ratio between relevant hits, missed events, and false alarms in real-world operation. A system can be highly sensitive on paper yet offer little benefit if it constantly reports harmless situations. Every warning demands attention. Too many false alarms lead to alarm fatigue: employees react more slowly, ignore alerts, or develop informal workarounds.

The costs of false alarms are manifold. They include lost working time in control centers, unnecessary deployments of security forces, potential operational disruptions, and a decline in trust in the technology. Conversely, a system set too low can miss genuine hazards. Optimization is therefore not a one-time technical procedure, but an ongoing operational process. Thresholds must be adjusted according to location, time of day, weather conditions, and type of event. A crowded platform on a public holiday presents different challenges than a depot at night.

For the tender process, this means that performance claims should be tested using real-world scenarios. Suppliers must not only demonstrate that their models can recognize objects, but also prove how reliably they operate under Moroccan lighting, climate, and traffic conditions. Dust, heat, backlighting, varying camera angles, and high pedestrian density all affect performance. Therefore, a pilot project at several typical locations is more economically valuable than a selection based solely on manufacturer specifications.

A key performance indicator (KPI) system that combines technical and operational metrics would be beneficial. This includes the number of relevant alarms, the average testing time, the false alarm rate, detected incidents, avoided downtime minutes, and the overall system availability. Only this combination allows for a determination of whether the investment is truly productive. A high detection rate without improved response time would merely be a technical success, but not yet an economic one.

Data protection as a matter of trust

The tender documents, as they currently stand, do not include facial recognition. This limitation is strategically significant. Facial recognition would make the project considerably more complex from a legal, ethical, and technical perspective. For the planned main applications, the identity of individual persons is not essential. The system must recognize that someone is in a restricted area or that a crowd has reached a critical density. It does not need to know who that person is.

Foregoing biometric identification can foster acceptance and support the principle of data minimization. Nevertheless, both traditional and AI-powered video surveillance remain an intrusion into privacy. Morocco has a legal framework for the protection of personal data in Law 09-08; the National Data Protection Commission (CNDP) oversees its implementation. For ONCF, this necessitates clear purposes, regulated retention periods, controlled access, and transparent accountability.

Economically, data protection is not merely an additional expense. Unclear regulations can delay projects, generate resistance, and force later modifications. Integrating data protection and information security into the architecture from the outset reduces long-term compliance and reputational risks. This includes separating live analytics from archiving, implementing automatic deletion concepts, logging access, encryption, and restricting access to authorized personnel only.

Purpose limitation is particularly important. A platform initially designed for intrusion and threat detection can technically be used later for more extensive behavioral analysis. This gradual expansion of its use carries political and societal risks. Therefore, new functions should not be legitimized solely on the basis of technical feasibility. Every expansion requires a separate benefit-risk assessment, a legal basis, and clear approval. Explicitly refraining from facial recognition is thus a sensible starting point, but only credible if it is secured by technical and organizational safeguards.

Cybersecurity is becoming a core task

With every networked camera, the digital attack surface grows. Cameras, local computers, network components, storage, interfaces, and central platforms can all contain vulnerabilities. An attack could intercept recordings, disable devices, suppress alarms, or generate false reports. Particularly critical would be manipulation that goes unnoticed. While a complete system failure is usually detected quickly, an undetected change to data or models can lead to permanently incorrect decisions.

The security architecture must therefore protect confidentiality, availability, and integrity equally. It is not enough to simply secure video data against unauthorized access. ONCF must also ensure that cameras are authentic, software versions are verified, models are not altered without detection, and alarms originate from trusted sources. Network segmentation is crucial in this regard: A compromised endpoint must not be able to grant direct access to mission-critical control systems.

Another risk lies in the supply chain. A system of this scale will likely include components from multiple manufacturers. Updates, remote maintenance, and cloud services can create dependencies that extend beyond the original contract term. Therefore, tenders should require security updates throughout the entire lifecycle, transparent vulnerability management processes, local disaster recovery capabilities, and clearly defined access rights. The ability to replace individual components without replacing the entire platform is also a key factor for both security and cost-effectiveness.

Cyber ​​resilience incurs additional costs but is not an optional quality level. The more security decisions are automated and centralized, the greater the potential damage from digital manipulation. Therefore, this project should not be treated as simply procuring cameras with IT security added later. It is a cyber-physical system in which digital decisions can impact real-world traffic flow.

Open architecture against vendor lock-in

The most strategically important procurement question is whether ONCF chooses an open, modular platform or a closed, all-in-one system. A proprietary solution can be implemented more quickly initially because hardware, software, and support come from a single source. However, high switching costs can arise in the long run. If cameras, analysis models, storage, and user interfaces are tightly tied to one vendor, any subsequent replacement becomes expensive and risky.

With a time horizon extending to 2030 and a significantly longer expected operating lifespan, technological flexibility is particularly important. AI models are evolving rapidly, while railway and camera infrastructure are often used for ten years or more. The architecture should therefore allow for updating or replacing analysis components without rebuilding the entire system. Standardized interfaces, documented data formats, and clear ownership rights to configurations and operational data are crucial for this.

Openness, however, does not mean combining as many providers as possible without central responsibility. Excessive fragmentation increases integration risks and makes troubleshooting more difficult. A clearly defined platform architecture with binding standards but competitive modules is economically sound. ONCF should retain control over the data model, interfaces, and security rules, while specialized providers can deliver individual functions.

Exit costs must also be included in the cost-benefit analysis. The lowest purchase price can become expensive over the product lifecycle if licenses, storage, maintenance, or upgrades are only available under monopolistic conditions. Therefore, the total operating costs should be evaluated over at least ten years. This includes not only hardware and software, but also energy, data transfer, training, product updates, cybersecurity, spare parts, support, and migration. A sound concept must make these hidden costs visible.

Opportunities for Morocco's digital economy

The project aligns with Morocco's national strategy, "Maroc IA 2030," which positions artificial intelligence as a driver of growth and sovereignty. The country aims for an additional economic contribution of 100 billion dirhams from AI by 2030, the creation of 50,000 direct and indirect jobs, and the training or certification of 200,000 people in AI skills. Such goals are ambitious and not automatically achievable. However, infrastructure projects like the ONCF system can act as concrete drivers of demand if local value creation is systematically considered.

The greatest local leverage lies not in the production of standard cameras. It lies in system integration, data engineering, model adaptation, cybersecurity, maintenance, and operational process design. Moroccan universities, technology companies, and startups could work on locally relevant datasets, testing procedures, and applications. This would generate knowledge that could later be transferred to ports, airports, industrial plants, logistics centers, and urban transportation systems.

A prerequisite is a procurement process that doesn't reduce technology transfer to non-binding letters of intent. Contracts can include training, local support capabilities, joint development, documentation obligations, and phased transfer of expertise. Equally important is access to anonymized or controlled training and test data. Without data access, local companies are often limited to basic installation and maintenance services, while the more valuable software and analytics capabilities remain abroad.

Morocco can also use the project as a benchmark for the African market. Many countries face similar challenges: growing cities, increasing traffic demands, limited security resources, and heterogeneous existing infrastructure. A modular, climate-resilient, and cost-effective solution could create exportable expertise. However, the ONCF project must deliver demonstrable results. A prestigious system without transparent performance indicators would be less valuable as a benchmark than a smaller, measurably effective approach.

Productivity instead of staff reduction

AI surveillance is often associated with savings in security personnel. This expectation is only partially realistic. A system with 6,000 cameras and numerous automated alerts still requires people to assess events, trigger actions, maintain technology, and monitor models. The economic benefit lies more in increased productivity than in simply reducing staff.

Traditional video surveillance scales poorly because humans can only attentively monitor a limited number of images. AI can pre-sort large volumes of footage and flag conspicuous situations. This shifts the focus from continuous observation to qualified evaluation. This can increase the effectiveness of existing teams and limit the additional staffing needs that would arise from network expansion and increasing passenger numbers.

At the same time, new job profiles are emerging. Control center analysts, data and AI specialists, cybersecurity professionals, system integrators, and technical maintenance teams are in demand. Existing employees need training to understand the limitations of the systems. Treating an AI warning as objective truth can lead to decisions that are just as dangerous as those who ignore all alerts. Human oversight, therefore, means more than just formal confirmation. It requires competence, time, and clear decision-making authority.

Work organization and technology must be designed together. Introducing new applications without adapting processes, shift models, and responsibilities increases the workload. Especially during the implementation phase, running legacy and new systems in parallel can create additional work. Therefore, cost-benefit analysis should consider transition costs and learning curves. Productivity gains rarely occur on the first day of operation; they arise through calibration, training, and organizational routine.

Financing and balance sheet reality

The investment drive is encountering a financial structure already heavily reliant on infrastructure. At the end of 2024, ONCF reported financial debt of approximately 36.5 billion dirhams, with around 71 percent attributable to infrastructure. Simultaneously, the company generated EBITDA of 1.949 billion dirhams in 2024 and a self-financing capacity of approximately 501 million dirhams. The net loss for the year was roughly 1.373 billion dirhams, primarily due to depreciation and infrastructure financing costs. According to ONCF, the result would have been positive without the infrastructure capital costs.

These figures illustrate a typical problem with state-owned railway systems: While operations may be economically efficient, the capital-intensive infrastructure generates high accounting and financial burdens. The 96 billion dirham program therefore cannot be financed solely from current surpluses. Government capital injections, loans from international financial institutions, and long-term financing structures remain crucial.

This necessitates rigorous lifecycle planning for the monitoring system. Initial hardware procurement is only one component of the costs. Ongoing expenses for licenses, storage, connectivity, energy, maintenance, and updates can reach a substantial level over the years. If these operating costs are underestimated during the planning phase, the result will be a digital infrastructure whose functionality is limited due to budget constraints. This would be particularly problematic because security platforms require continuous maintenance and cannot remain largely unchanged after completion like a one-time construction project.

At the same time, the project should not be evaluated solely based on direct revenue. Security investments generate their benefits primarily through avoided damage, more stable operations, and improved utilization. A robust cost-benefit analysis must therefore use scenarios: How many relevant incidents occur, how much faster are they detected, what costs due to delays or damage can be avoided, and how do insurance, personnel, or maintenance costs change? Without such assumptions, the economic evaluation remains too abstract.

From study to robust implementation

The four-month planning phase should conclude with a clear prioritization. Not all 6,000 cameras and not all AI functions need to be implemented simultaneously. A phased approach reduces technological and financial risks. Initially, locations should be selected where security needs are high, data availability is good, and the benefits are measurable. These could include busy train stations, selected high-speed rail sections, depots, and known access points.

During a pilot phase, models must be tested under real-world conditions. Different seasons, times of day, weather conditions, and passenger densities are crucial. Following this, ONCF should only scale those applications that meet agreed-upon minimum performance standards. Features with too many false alarms can be further developed or discarded. This discipline prevents the common mistake of purchasing a broad feature set whose practical performance falls short of the promised presentations.

The tender should also establish a robust data and governance structure. This includes responsibilities for data quality, the release of new models, security updates, data protection, incident analysis, and performance monitoring. Every AI function needs a business owner within the operation. Pure IT responsibility is insufficient because the consequences of decisions impact safety and traffic management.

Finally, the system requires independent acceptance testing. Manufacturers should not define and evaluate their own performance. Test data sets, scenarios, and key performance indicators must be controlled by ONCF or verified by independent bodies. It is crucial not to consider only average values. A system can perform well overall but fail precisely in critical situations, such as strong backlighting or high crowd density. Acceptance and ongoing monitoring must therefore be risk-based.

Morocco's rail network: Why an intelligent system architecture is more important than pure camera technology

The planned AI and hypervision strategy is neither a mere camera project nor proof that Morocco's railway will be fully automated in the future. It is the preparation for a new digital layer above the growing rail network. Its potential lies in faster response times, better coordinated safety processes, greater operational stability, and the development of transferable technological expertise.

Economic success depends less on the number of cameras than on the quality of the architecture. 6,000 devices and 1,800 AI-enabled units are achievable targets, but not success metrics. Crucial factors are avoided operational disruptions, short response times, manageable false alarm rates, high system availability, and low lifecycle costs. Equally important are data protection, cyber resilience, and the ability to switch providers or models later.

From an industrial policy perspective, the project offers the opportunity to build Moroccan expertise in computer vision, systems integration, and critical infrastructure. However, this opportunity will only be realized if open standards, local knowledge transfer, and measurable participation become binding. Otherwise, Morocco will remain a buyer of an imported platform and bear the costs of technological dependence for years to come.

The rationale is therefore: ONCF is right to plan the system architecture before large-scale procurement and to forgo facial recognition for the time being. However, the next step must go beyond a technical blueprint. What is needed is an economic and organizational operating model that makes benefits, risks, responsibilities, and follow-up costs transparent. If this succeeds, the small tender could become a key project for a safer, more efficient, and digitally sovereign rail transport system. If it fails, there is a risk of an expensive network of cameras that generates large amounts of data but provides too few reliable decision-making tools.

I would be happy to serve as your personal advisor.

You can contact me at wolfenstein∂xpert.digital or

Just call me on +49 7348 4088 965 .

LinkedIn