The Center for Data Innovation recently spoke with Csaba Sandor, CEO of Surviot, a Hungary-based company that is solving data integration challenges in the construction industry. Csaba explained how Surviot combines sensor data with machine learning to enhance structural monitoring, improve safety, and boost efficiency on construction sites.
David Kertai: What inspired the creation of Surviot?
Csaba Sandor: Surviot grew out of a real-world challenge during a project to construct a tunnel in Sopron, Hungary. A friend of mine, who had come from Norway to lead the surveying team, couldn’t find a platform that could integrate data from different sensors, measurement tools, and geolocation and geotechnical sources. Most existing platforms didn’t allow teams to combine data from different types of sensors or tools—each system was tied to a specific equipment supplier. If you used one company’s platform, you had to use their entire suite of hardware, whether it fit your project needs or not. This kind of lock-in made it difficult to integrate data across sources, created inefficiencies, and left blind spots in structural monitoring. We decided to build a platform that could fuse data from multiple sources, visualize it for users, and generate shareable reports for stakeholders. What began as a solution for a single project quickly revealed broader industry potential, and we’ve since scaled it to support a wide range of future construction projects.
Kertai: How does Surviot integrate and analyze diverse sensor data for structural health assessments?
Csaba: Surviot helps construction teams monitor the condition of physical structures—like bridges, tunnels, and buildings—by collecting and analyzing data from a wide range of sensors. We follow a four-step process to make this structural health monitoring seamless, flexible, and actionable.
First, we collect data through three methods. Our own data acquisition tool connects directly to various sensors and transmits real-time data to the cloud. We also retrieve data from third-party platforms and APIs, ensuring compatibility with different tools. And when teams take manual measurements on site, they can upload the data directly to the platform.
Second, we analyze the data using an open engineering platform that supports modeling, simulations, and client-specific machine learning and AI modules. These tools detect patterns, correlations, and anomalies, transforming raw sensor data into meaningful insights.
Third, we visualize the data through dashboards, graphs, and 3D models. These help spot issues early and make quick, informed decisions on-site.
Finally, we generate real-time alerts and customizable reports to track changes over time. Project teams and stakeholders can share these outputs to support coordinated, data-driven decisions and maintain safety throughout the construction process.
Kertai: What steps do you take to ensure the accuracy and reliability of sensor data?
Csaba: We rely on machine learning to validate and filter incoming sensor data in real-time. Our models cross-reference information from multiple sources—including environmental conditions—to isolate what really matters. For instance, we separate out the effects of temperature on structural movement, which is often crucial for assessing a structure’s true behavior and long-term sustainability. By adding this layer of context, we help users make decisions based on insights they can trust.
Kertai: What machine learning methods does Surviot use?
Csaba: We build our machine learning modules to detect anomalies like sudden structural shifts or deterioration and to recommend appropriate actions based on those insights. We also clean the data by filtering out noise and irrelevant signals that could distort analysis. We primarily use time-series analysis, anomaly detection, and sensor fusion techniques, which we implement with Python-based machine learning libraries and frameworks.
Kertai: How is Surviot evolving with AI, and what’s next for its role in construction?
Csaba: We continue to make AI and machine learning core to Surviot’s evolution. These tools help us move beyond simply collecting and displaying data, we now interpret it in real-time, identify trends, and predict risks. As we expand beyond Hungary into Eastern and Southern Europe, Central Asia, and India, we’re focusing on making these tools practical and accessible for everyday use in underserved construction markets. Our goal is to earn the industry’s trust by showing how our technology can directly improve project outcomes. Looking ahead, we plan to expand deeper into the EU, particularly in Germany, by refining our machine learning capabilities to meet the growing and changing demands of the construction sector.
