Digital transformation? Not without uniform data models

For years, real estate has been under pressure to reinvent itself digitally, with varying success. Gerald Kremer argues that a common data model is the key to change.

The real estate industry has been undergoing continuous digital transformation for a decade, but progress has been slower than many expected. Individual processes have become more digital, and the first AI applications are finding their way into the sector.

Nevertheless, much remains fragmented, inefficient and manually controlled. The fundamental cause: there is a lack of uniform data models across the industry that enable the consistent exchange and automation of information.

Unstructured data slows efficiency

Data is at the heart of every digital strategy, yet in the real estate industry, it is often still unstructured and unevenly distributed. Many companies, including large asset managers, now collect huge amounts of data from a wide variety of sources. But consistency remains a problem: even with identical service providers in different countries, data structures vary considerably.

Unlike large providers in the securities sector, globally active property owners are dependent on the respective property managers. The weaknesses are particularly evident when a property changes hands: the handover mechanism is usually still analogue and requires a lot of manual work.

The inconsistent database prevents digital tools from reaching their full potential. This is because machine learning models (AI) need to be trained. And this requires a larger and higher-quality training data set, which many companies, especially smaller ones, do not have. A public data platform would be helpful so that all players can train their AI and put the models into productive use in their companies. Without a binding standard, the real estate industry also runs the risk of remaining stuck in isolated solutions and only achieving fragmentary efficiency gains.

First steps towards a uniform database

Large asset managers are already investing heavily in data models and the harmonisation of their data stocks. Admittedly, the effort involved is considerable. Studies show that around 90% of the resources available for digital development are spent on preparing and ensuring data quality, with only 10% going towards the development of specific AI solutions. However, the investment is worthwhile, as valid data analyses form the basis for better decisions on rental agreements, location assessments and investment decisions.

So why is digitalisation still progressing so slowly? On the one hand, because many digital projects have not delivered the hoped-for immediate 100% improvement. According to the study The GenAI Divide: State of AI in Business 2025 by the Massachusetts Institute of Technology, only about 5% of AI pilot programs have a positive impact on the turnover of the respective companies. The way in which companies introduce AI is crucial: according to the study, the purchase of AI tools and the establishment of partnerships are successful in around 67% of cases, whereas internal solutions are only successful in a third of cases.

On the other hand, the industry has been able to work well with analogue processes so far, which has reduced the pressure to innovate. Digital transformation therefore needs advocates in the boardroom who will push it forward with determination and provide the necessary budgets.

Looking to the future

One thing is certain: uniform data standards could give a significant boost to the currently slow pace of digitisation in the real estate industry. Data storage must then become platform-based and interoperable. Only then can gaps in the process chains, for example when owners change, be closed and machine learning models can really add value. For real estate of the future, this means networked, digital value creation that is not only more efficient, but also more transparent and agile. l

Gerald Kremer is a member of the management board (COO) at Union Investment Real Estate