AI as a Compass: How a Cantonal Authority Turns Data into Safety

The authority has set itself an ambitious goal: fewer accidents, less suffering, and reduced economic damage. Its lever of choice: modern technology. As is often the case, it all began with a simple question: Can AI help predict, reduce—or even prevent—accidents? 

The answer is now a platform that translates data, forecasts, and imagery into preventive impact. 

As the orchestrator of infrastructure, mobility, and safety, the cantonal authority is far more than a mere administrative body. Its requirements for data-driven innovation were accordingly high: transparency, data protection, scalability. At the same time, it lacked the expertise and resources to implement such a transformation on its own. 

So, at the end of 2022, a project was launched with the ambition of delivering more than just a technical solution. The goal: to intelligently connect existing data sources—from weather and accident data to photogrammetry. The result would be a platform capable of generating insights: Where do accidents cluster? Which structural conditions contribute to them? Which countermeasures are most effective?

The authority already had a wealth of information at hand: weather and accident statistics, police reports, images. But they were scattered, siloed, and hard to access. Tools such as IBM SPSS for statistical analysis, Geographic Information Systems (GIS), or Spatial Information Systems (SIS) were available, but there was no way to make practical use of the isolated datasets. 

The central office for accident prevention had neither access to modern forecasting models nor dedicated AI staff. What was missing was not high-quality data—but a unifying operating system to harness it and extract actionable insights. As part of a strategic inquiry into the “feasibility of AI in accident prevention,” the TIMETOACT GROUP came into play. Step by step, this developed into a partnership built on strategy, substance, and scalability. 

The transformation began with a structured AI Strategy Workshop. In this early phase, data availability, organizational maturity, and technical infrastructure were assessed—with one clear conclusion: strong data potential, but low internal maturity in terms of AI readiness. 

The outcome was the architecture of a scalable, Azure-based data and AI platform. It integrates Synapse, Machine Learning, OpenAI models, Power BI, and GIS tools—within a managed Azure tenant operated by X-INTEGRATE. At its core lie a Data Lakehouse and a flexible access layer that link structured and unstructured data, making them available for visualization and model operation. 

So far, three use cases have been implemented: 

• A central accident dashboard analyzing traffic accidents by region, vehicle type, time, and weather conditions. 

• A computer vision solution analyzing entry and exit angles at roundabouts—with concrete recommendations for structural improvements. 

• An accident prediction model with 7-, 14-, and 21-day windows that combines weather, traffic flow, and road conditions to forecast potential accident hotspots. 

What once took weeks can now be done in a fraction of the time. For the first time, the authority has a holistic view of traffic activity across the canton. Forecasts are significantly more accurate than previous models, and recommendations are more robust. 

The platform now provides a sound basis for decisions on infrastructure measures, emergency service planning, and strategic investments. It is modular, technology-agnostic, and built for the future. While day-to-day operations are managed by external partners, the authority has begun building its own AI competence structures. 

The cantonal traffic authority demonstrates how modern public administration can operate: data-driven and purpose-driven. With a clear focus on safety, prevention, and efficiency, it has built a technical infrastructure that not only protects lives in the long term but also reduces operational workload. By investing in internal competencies and modular platform design, the system is equipped to remain successful well into the future.