Agentic AI in the Operations Center: A Glimpse into the Future of Rail Dispatching with SBB

What will the dispatcher’s workplace look like in the future? As part of the SBB AI Challenge, adesso teamed up with Swiss Federal Railways to develop a “concept car” — or rather, a “concept train” — for incident and disruption management. Here’s a look at this exciting journey between human expertise and artificial intelligence.

The SBB's operations centers are at the core of Switzerland's rail network. When a disruption occurs—whether it’s a door malfunction or a weather-related outage—dispatchers must make decisions within seconds. They coordinate emergency crews, reroute trains, organize replacement transport, and inform passengers. The central communication tool for this is the ALEA system. It is highly functional and has been continuously adapted to the needs of all user groups. As a result, it has become a platform that connects various SBB departments as well as other transportation companies throughout Switzerland—enabling comprehensive incident management. But its origins date back many years, which is why we asked ourselves:

How could state-of-the-art technology better support people in these challenging situations if we were to redesign ALEA from the ground up?

At adesso, we approached this core question with great passion as part of the AI Challenge.

Instead of merely “optimizing” the existing system in isolated areas, we chose the “Concept Train” approach. We wanted to holistically demonstrate the possibilities offered by agentic AI in the complex world of rail logistics.

Together with experts from SBB, we developed a working prototype that demonstrates how human-system interaction could change.

Our focus was on three pillars:

1. The “listening” system (ChatOps)

Much of the communication during disruptions takes place via phone and radio. In our approach, this spoken information is made directly usable through AI-powered transcription. The goal: While the dispatcher is still on the phone, the system is already preparing the ALEA case in the background. Message intake and case creation happen in a single step—a massive time saver.

2. Intelligent Assistance

Similar to modern driver assistance systems, the AI checks inputs for plausibility and suggests appropriate phrasing based on the current context. This improves the quality of the data for the respective events, which in turn enhances the quality of the suggestions the system generates.[Text Wrapping Break]Furthermore, such an assistance system has another positive effect: it supports dispatchers in formulating reports, actions, or internal messages. While this may sound trivial, it can save valuable minutes in everyday work, because bullet points are sufficient for creating content, and the message is enriched with additional relevant information.

3. Automated processing of routine tasks

Routine tasks, such as summarizing incident timelines for reporting, are handled by AI, allowing dispatchers to focus on solving the problem. The “overhead” associated with tedious but necessary work is significantly reduced, giving teams in the operations center more time for what matters most: ensuring passengers reach their destinations on time.

What makes this project particularly valuable is the open and constructive exchange with SBB. Feedback from SBB has confirmed that speech-to-text is a promising field that can deliver immediate added value for incident management. The vision of creating cases more quickly via voice input and relieving dispatchers through assisted logging was validated as a practical and forward-looking approach.

At the same time, the joint analysis has shown where the challenges of real-world implementation lie. Issues such as noise levels in the operations center during voice control or the availability of networked data (e.g., rolling stock and personnel) are critical factors that must be taken into account in further development. It is precisely these insights — the balance between technical vision and operational feasibility — that define the value of such a prototype.

We are particularly pleased by SBB’s recognition of our work and especially by the emphasis on “constructive collaboration and openness in communication.” For us at adesso, this is the greatest compliment. At the same time, it spurs us on to continue living up to our mission: to understand complex domains and improve them with technological innovations such as Agentic AI.

Even though a comprehensive rollout of such a system requires investment and adjustments, together we have outlined a vision of what the next step toward the future of dispatching could look like: more efficient, more intuitive, and more supportive of the dispatcher in control.

We would like to thank SBB for their trust and the opportunity to work together on the mobility of tomorrow!

From the “Concept Train” to reality: Setting the course for the control center of the future together

Building on our successful collaboration with SBB, we invite you to join us in validating how Agentic AI can boost the efficiency of your control center while making technical complexity manageable. From an operational perspective, our approach offers a clear vision of how existing systems and processes can be future-proofed through intelligent AI assistance. Day-to-day operations benefit from a noticeable reduction in the workload of dispatchers, as agent-based AI takes over routine tasks such as reporting and helps ensure data quality during split-second decisions through automated plausibility checks and well-founded recommendations for action.

In doing so, we specifically address critical success factors such as the quality of passenger and incident information, the operational results of dispatch measures, and the efficiency of processes and documentation in the operations center. Let’s explore how we can reduce administrative overhead through the targeted use of Agentic AI, so your dispatchers can fully focus on resolving disruptions in stressful situations. Together, we’ll transform this technological vision into a robust, productive solution for the dispatch operations of the future!