BIM in vacant buildings: is it worth it?

An empty building initially seems to be the simplest starting point for a conversion. No occupants, few operational restrictions and unrestricted access.  Where do pipes actually run? Which walls have been changed in recent decades? Which documents still correspond to the built condition?

These questions are gaining importance. In the seven largest office property markets of Berlin, Düsseldorf, Frankfurt am Main, Hamburg, Cologne, Munich, and Stuttgart, approximately 8.5 million square meters of office space were vacant in mid-2026. The average vacancy rate reached 8.5 percent, up from 7.7 percent a year earlier. At the same time, the gap between modern, well-equipped buildings and properties in need of renovation is growing. Alternative uses such as education, healthcare, hospitality, or residential are therefore becoming more interesting for existing property owners.

Improving the existing building stock is also crucial for climate action. In 2025, the building sector in Germany caused 103.4 million tons of CO₂ equivalents, 3.4 million tons more than in the previous year. The current projections of the Federal Environment Agency still show a significant need for action by 2030. Those who want to develop buildings further instead of replacing them therefore need reliable information about what is available. BIM can provide a basis for this. However, the first step is a survey of the building, before any modelling begins.

First capture, then model

The Federal Chamber of Architects (BAK) completely revised its guide "Digitalization and Construction in Existing Buildings (opens in a new window)" in 2025. Digital surveys of existing buildings plays a central role in it. Depending on the project, methods such as tacheometry, laser scanning, drone imagery, and other digital measurement procedures are used. The BAK also points out an important difference: a point cloud or a laser scan is not yet a BIM model. The captured geometry must then be structured and modeled for the respective purpose.

How complex this becomes depends on the project. For an initial feasibility study, floors, rooms, facades, structural framework, and essential access might suffice. In a complex renovation of the building services, shafts, pipe routes, and technical systems must be recorded in much greater detail.

Therefore, it should be clear before the first scan what decisions are to be made later with the model. The BAK explicitly recommends defining the requirements for surveys of existing buildings early between the client, planning, and surveying. Not everything that can be technically captured needs to be modeled. Especially in vacant buildings, the conditions for this are comparatively good. Rooms are more easily accessible, measurements can be organized without ongoing user operation, and missing areas can be more easily recorded later. While this does not save the actual modeling work, it does facilitate data collection.

Scan-to-BIM becomes more automated

At this point, technology is developing rapidly. The classic Scan-to-BIM process is still labor-intensive: point clouds initially contain measurement points. Whether a wall, a window, a pipe, or a staircase is hidden behind them must be recognized and then converted into an object in a model. Fraunhofer IPM (opens in a new window) is now working on automated procedures that partially take over these steps. Cameras, laser scanners, and other sensors capture interiors and facades. AI procedures classify objects directly in the point clouds and can, for example, recognize doors, windows, pipes, or stairs. BIM-compatible models can be derived from this.

This approach is also being further developed in serial retrofit. In the SerialBIM research project, launched in 2026, two multi-family houses in need of renovation are to be digitally surveyed. Among other things, AI-based digital surveys of existing buildings, automated model generation, and the evaluation of the accuracy of the generated building models are being investigated. The project runs until 2028.

A precise model is not automatically complete

The direction is thus clear: in the future, there should be less manual modeling work between the scan and the finished model. However, for complex existing buildings, this does not yet work fully automatically. Especially unusual constructions, hidden components, and incomplete existing documentation still require expert interpretation. This is an important limitation of digital surveys of existing buildings. A laser scanner can capture visible geometry very precisely. However, it does not automatically see what is behind a wall, above a closed ceiling, or in an inaccessible shaft. A cleanly modeled building can therefore suggest an accuracy that is not present for individual areas. In existing buildings, known geometries, adopted plan information, and assumptions should therefore remain distinguishable from each other.

The Federal Chamber of Architects dedicates a separate topic to modeling accuracy and information density in its current guide. The necessary level of detail depends on the planning task. BIM Deutschland also emphasizes for digital twins that a digital model does not have to be as complete and detailed as possible. Rather, the necessary information depth results from the specific application. This is crucial for economic efficiency. Anyone who has every socket, every pipe, and every material of a vacant building recorded, even though the primary goal is only to check whether an office property is fundamentally suitable for a new use, primarily generates data costs.

First check conversion variants digitally

Conversely, a properly tailored existing building model can provide early answers. Is the existing room structure suitable for a different use? Which components can be retained? Where do new lines clash with the existing structure? Does an existing technical route actually need to be relocated, or can the plan be adjusted?

Especially with a change of use, many requirements intertwine. Floor plans change, fire protection and access must be re-evaluated, building services must meet different performance requirements, and additional shafts or installations may be added. Variants can be compared in the model before components are removed or new openings are created. The value is not that every decision is made automatically. Rather, planners receive a common spatial basis on which effects become visible earlier.

Digital documentation is also becoming more important for handling existing materials. The Federal Chamber of Architects has included sustainability and circular economy in the current new edition of its guide to existing buildings. Information about which components and materials are present can help to consider preservation, reuse, and deconstruction already in the planning stage.

From existing model to digital twin

After renovation, the data can continue to be used. A structured model can provide information for space management, maintenance, or energy analyses. If system information, manufacturer data, and maintenance specifications are cleanly added during the conversion, the operator does not have to collect this data again later. The next step is the connection with current operational information. Fraunhofer ISE, for example, is working on digitally capturing existing information from building services systems and connecting it with BIM and time series data from building automation. This allows system statuses to be monitored and operating strategies optimized.

BIM Deutschland has also increasingly addressed this topic in 2026. A dialogue forum on digital twins in existing buildings focused on 3D laser scanning, existing plans and schematics, system information, measurement data, and interfaces to building automation, among other things. From these different sources, a digital representation can gradually emerge, which is used for monitoring, optimization, or predictive maintenance. The term "digital twin" should not be used too early. A scanned and modeled existing building is initially a digital model of the existing building. Only when the model is connected with current information from the real building and updated for specific operational processes does the additional benefit of a digital twin arise.

Not every vacant building needs the most detailed model

Thus, the question "Is BIM worthwhile for vacant buildings?" does not lead to a blanket answer. For a building that is to be completely deconstructed in the short term, extensive modelling may be disproportionate. For a complex revitalization with several usage variants, densely packed building services, or planned long-term use, the calculation looks different.

Digitization should therefore not begin with the question of how detailed a building can be scanned. What matters is what should happen with it afterwards. Which planning risks should be reduced? Which variants should be investigated? Which existing parts should be preserved? Which information will be needed later during operation? The clearer these questions are answered, the more precisely it can be determined what needs to be recorded and modeled. The technology for this is developing rapidly: mobile measurement systems, automated point cloud processing, and AI are gradually reducing the effort between the building survey and the BIM model. At the same time, current research projects show that full automation still requires work.

Vacancy offers a special window of opportunity for this. The building is accessible before the next use begins and new installations change the existing structure again. Anyone who uses this phase to specifically digitally record the existing structure not only creates a better basis for conversion. They can also decide more precisely what can remain of the existing building and how the next use will arise from it.