Planon RE: reliable data beats AI in helping real estate investors make good decisions

Artificial intelligence cannot help real estate investors make good decisions unless it is backed by reliable data, delegates heard at the ‘Prevention vs Reaction: How Data Control Enables Better Investment Decisions’ webinar, organised by Real Asset Media in partnership with Planon Real Estate.

“Data quality is crucial”, said Daan Berghmans, Solutions Director, Planon Real Estate. “AI models do not operate well on poor data and cannot improve them. Put simply, if rubbish goes in then rubbish will come out, so it is very important to be in control of the data because AI on its own will not solve the problem.”

The ease of using an AI tool can generate a false sense of security: but AI is trained to satisfy the user and deal with the input that is provided, so if there are issues with the files dropped in an AI model the output will inevitably be compromised.

“I would not entrust operational decisions to AI now, because data definitions and structures are often skipped”, said Berghmans. “AI could hallucinate when insufficient data is provided or simply make a calculation mistake. If an investor does not watch out, AI will create a growing pile of Excel files, each of which has to be verified.”

Getting things right is particularly important in today’s market, as geopolitical uncertainty and low liquidity mean that investment decisions are more complex and take longer, while investors are more demanding.

“Now that success doesn’t come as easily as it used to, we’ve noticed that investors have become more critical and are asking for more frequent updates on the investments we make”, said Marlijn Hanna, Associate, Asset Management, Edmond de Rothschild. “Data is playing an increasingly important role in this process. It is essential to be able to quickly create accurate analyses and run different scenarios. Having a solid database is the foundation for all of this, because if you put poor-quality data into a system, you will never get high-quality analyses or reports out of the system, even with AI.”

The real estate market has long relied for calculations on Excel spreadsheets, which are often revised and updated and come in many different versions, which is time-consuming and makes costly mistakes more likely.

“We’ve found that relying on Excel spreadsheets is no longer enough”, said Hanna. “They are time-consuming and sensitive to errors. That’s why we started looking for a system where we could store and manage all of our data in one place. Having a single source of truth allows us to work more efficiently, improve accuracy, and provide investors with better and faster insights.”

In an ideal scenario, AI becomes the layer that helps clean, structure, and activate data within a trusted system. The right software can create an agent framework that allows AI to interact directly with the database, extracting and transforming structured or unstructured data based on predefined data definitions. Users can then review the data, see what has changed, and understand the impact before approving it. Once approved, the data is imported into a validated financial model, where it can be further analysed and adjusted, by using custom interfaces that are optimised for the task or by interacting with AI through natural language.

“I believe AI will mainly have an impact on the reporting side of our job, which is extremely important for investors, but it is also very time-consuming and requires a significant amount of manual work”, said Hanna. “I think AI will help automate the process of transforming data into the appropriate reporting formats and models. As a result, asset managers will be able to spend more time reviewing the information and focusing on decision-making rather than preparing reports.”

AI could also help compare data from the previous quarter with the current quarter and identify any unusual changes or inconsistencies, which would provide an additional layer of validation and help determine whether changes in the data actually make sense.

“Everybody says that AI is going to make their work easier, and the use of AI is something that cannot be stopped”, said Berghmans. “However, financial forecasting and decision making requires a large context window, and AI models cannot support this for a complete portfolio. Therefore, I think that portfolio management and portfolio risk analysis will still be done by people and companies. It is important that you should always be in control, not the AI tool.”