AI hype: Lack of data quality is becoming an investment risk in residential construction

From the point of view of ProOmea GmbH, one of the biggest bottlenecks in the current market lies in fragmented data, manual review processes and incomplete decision-making bases. These weaknesses not only limit the speed of classic purchase checks, but also the reliability of AI applications.

This is because generative AI can only take into account information that is available in a structured and sufficiently high quality. If important information is missing, for example on building law, eligibility for funding, construction costs, rental rates or exploitation scenarios, a complete decision-making model is not created. The AI then fills the data gaps with statistically plausible assumptions. This results in an answer that sounds convincing, but is based on an incomplete and possibly wrong basis for decision-making. The greatest danger therefore lies not in the lack of an answer, but in a convincing answer based on insufficient data.

This is exactly where ProOmea, a Stuttgart-based company with its Decision Intelligence Platform, comes in: It brings together market data, funding structures, project parameters and a large amount of real transaction data in a structured analysis process. This creates a reliable basis for purchase and investment decisions in residential construction. An in-house IT team is continuously developing the digital solutions for this purpose. In recent years, ProOmea has accompanied the realization of more than 1,000 apartments in Germany.

"The real bottleneck in the current market is not the lack of technology, but the fragmentation and quality of the data structures," says Jeremy Justin Jahn, Managing Partner of ProOmea GmbH. "Many market participants still check real estate investments with a high level of manual effort, inconsistent data sources and classic Excel models. In a market environment with volatile construction costs, complex subsidy programs and high demands from institutional investors, this way of working is increasingly becoming a risk."

This approach is particularly relevant in affordable housing. Here, the rental model, eligibility for funding, construction costs, regulatory requirements and return expectations must be compatible with each other at an early stage of the project. Many projects are socially sensible and politically desirable, but not yet structured in such a way that investors can make a reliable purchase decision.

"An AI works with probabilities. In the end, an investment decision is binary: to do or not to do," says Jahn. "You can't buy 80 percent of a property. Therefore, in addition to technology, mathematics, market knowledge, funding expertise and experience from real transactions are needed in the structuring of complex projects .“

ProOmea's Decision Intelligence Platform bundles several levels of analysis and testing. This includes market and rental data, location assessments, project parameters, funding programmes, regulatory requirements and economic scenarios. The aim is to create a reliable overall picture from a large amount of individual information.

In this way, it can be seen at an earlier stage whether a project is economically viable in its existing form or whether other use, funding or exploitation scenarios would make more sense. Especially in residential construction projects, this early view can be decisive before central decisions such as the mix of space, floor plans, energy standards or eligibility for funding can hardly be changed.

From ProOmea's point of view, the final decision will remain with humans in the future. Technology can structure data, make risks visible and prepare usage scenarios. However, the responsibility for the use of third-party capital is not transferable.

"Real estate investments are a business of trust and liability," says Simon Bender, Technical Product Engineer at ProOmea. "It is therefore crucial that results remain comprehensible and that people recognize the data, assumptions and models on which they are based."

For ProOmea, the progress of data-based real estate analyses therefore does not lie in autonomous purchasing decisions, but in better decision-making bases. Those who combine market, location, subsidy and project data at an early stage can review residential construction projects more quickly, assess risks better and prepare investment decisions more robustly.