How AI is changing the real estate world

Detailed valuations, swifter transactions, more efficient use of energy: like in other sectors, artificial intelligence has the potential to transform the real estate world. Jobs in the industry will also be impacted by the resulting upheaval. By Claus Hornung

The investor was satisfied. Four brokers had been given the task of finding real estate in the US city of Denver that would meet his expectations. The brief was to find buildings similar to those the client had previously acquired. In the end, there was a clear winner with exactly the right price, location and facilities. We don’t know whether the three brokers whose proposals were rejected were disappointed, but it’s a fact that the winner showed no human emotion at his success, because he wasn’t human – the winning proposal was selected by artificial intelligence.

This example from the US shows that artificial intelligence (AI) has the potential to transform the real estate industry. “This technology will have an impact on every single activity in the real estate industry – and on every phase in the life cycle of buildings,” says Heiko Gsell, Professor of Business Information Technology and Director of the Digitalisation and Real Estate Management programme at the EBZ Business School in Bochum, Germany. 

AI makes real estate processes significantly easier and more efficient

AI reduces energy consumption by using past consumption data to make reliable predictions about the future. It improves security because it can detect noises and camera images faster than security personnel, analyse the information, and alert the police or fire brigade in the event of danger. It can identify patterns in the operation of heating, ventilation and air conditioning systems and thus accurately forecast dates for repair and maintenance work. It also speeds up transactions by matching offers from previous acquisitions with property descriptions and then crafting customised purchase offers – as in the case of the Denver example cited above. “Processes become significantly easier and more efficient,” says Carsten Kreutze, Managing Director of Bonn-based software company Recogizer, which uses AI to reduce energy consumption in buildings. “Ongoing development will at some point make it possible to collate all the data relevant for real estate transactions and managing properties at the touch of a button,” he adds. What condition is the property in? When will the radiators need replacing? How many tenants are there, and when do their leases expire? Can I push through rent increases?

Real estate industry is a laggard in digitalisation

But none of this will work without good data quality and availability, notes Patrick Penn, CEO of Docunite, the provider of an AI-based document management system: “For a sale, for instance, you need not only leases and land register extracts but also the energy performance certificate or more exotic documents, such as information on contamination of the site by explosive ordnance.” In many companies, these documents are typically spread over a range of departments. “The leases might be held by the property manager, and the building services plans by the asset management team,” says Penn. Sometimes a search will reveal multiple versions of the same document with different dates: “It can take six months for everything to be sent to the buyer in some cases.”

The real estate industry is still too analogue in many respects, notes Penn. “This is an area where companies need to do their homework.” Which means digitalising data and merging both their own internal data and publicly available data via interfaces. “This will generate opportunities that can deliver a huge competitive advantage,” says Penn. That is particularly true since the inception of generative AI, like Chat GPT. “Now you can chat with your documents,” Penn adds.

After all, the user isn’t interested in greater ease of access to a document, what they want is the information contained in the document.
Patrick Penn Patrick Penn, CEO, Docunite

What this means is illustrated by a test conducted by his team. They fed Chat GPT with an energy performance certificate and then had a colleague with Iranian heritage ask what types of windows were installed in the building – in Farsi, the Iranian national language. “She received a correct answer,” Penn says with a smile, “in Farsi.” For him, this is the very definition of AI: “After all, the user isn’t interested in greater ease of access to a document, what they want is the information contained in the document.”

Property management is already significantly further ahead in its necessary pursuit of digitalisation, says Recogizer Managing Director Kreutze. This is due to the fact that digitalisation is a lever for achieving an important goal of many property companies: reducing energy consumption. “For years there was very little interest from owners and tenants,” says Kreutze, “but a lot has happened in the past two or three years. There’s pressure on the one hand from users, who want lower service charges and a greener property, and from owners, who want to operate their buildings in a more climate-friendly way as a means of reducing devaluation risks. At the same time, there is regulatory pressure – such as the requirement that the building stock should be carbon-neutral by 2045.”

Use of AI in property management can reduce energy consumption,

Many companies now have a kind of control room that brings together their air conditioning, ventilation and heating systems, says Kreutze: “But it’s only with AI that you can really address the big consumers.” This is because AI can take into account a host of factors that influence consumption: What will the weather be like over the next few days? Is the space let or vacant? Does a room have unusually large windows? What is the building’s function – is it used as a gym, an office or for retail? At what time would you normally expect the biggest concentration of people? 

AI not only recognises correlations, it can also use them to come up with answers to complex questions, says Carsten Kreutze: “A question of this sort might be something like why is the energy consumption in our top five properties higher than in the others.” Furthermore, artificial intelligence can make specific suggestions for improvement, advising on whether flow temperatures should be optimised, for example, fan speeds adjusted or system operating hours made dynamic. “With measures like these, we’re already achieving average savings of 28 percent in energy consumption in commercial real estate,” says Kreutze.

Only with AI can you really address the big consumers of energy in a building.
Carsten Kreutze Managing Director, Recogizer

Support for facility managers, new jobs for prompt engineers

A question that currently looms large in the real estate industry is whether use of AI will cost jobs. “In facility management, finding well-qualified employees is a major challenge. Facility staff also currently have to perform a large number of tasks that are often technically very demanding – frequently all at the same time,” notes Kreutze. 

AI can provide valuable support here, he says: “It enables facility managers to concentrate on key tasks and supports them with smart solutions in areas where in-depth technological expertise is sometimes essential.” Docunite CEO Penn likewise takes a nuanced view. While stressing that “employees who don’t embrace AI run the risk of becoming replaceable”, he also predicts that new job profiles will emerge at the same time. “Like other sectors, the real estate industry is going to need prompt engineers,” says Penn, “which means people who ask AI highly qualified questions to get it to deliver the best possible results. That will open up many opportunities for career changers.”

Business information specialist Gsell also cites other reasons why artificial intelligence should not be left to manage or sell buildings entirely autonomously in the future. “Even granted that systems continue to learn, we know that they often come to the wrong conclusions. So the final decision should always be made by human beings.” 

But safety is not the only factor to be considered, says Gsell: “I don’t think AI can recognise the potential for a property to be repurposed, for example. That’s why I believe human assessment will always play a role in valuations.” Maybe Denver was just a one-off.

By Claus Hornung

IoT: Overcoming the language barrier

IoT solutions for buildings are still beset by a Babylonian-style confusion of languages. The DROPS research project aims to change that. By Dagmar Hotze

Their sheer complexity in IT terms means that smart buildings are currently few and far between. And yet, they are considered an important building block for sustainable cities and districts. Consulting firm Drees & Sommer, construction services provider Strabag and HafenCity University have come together at the initiative of digital property manager Reos in a research project aimed at developing the required data standards. The project title is DROPS, which stands for Data Standards for Resource-Optimized Production and Service Processes in buildings and quarters.

The DROPS team is committed to data standards (from left to right): Wolfgang Willkomm (HafenCity University Hamburg), Phillip W. Goltermann (Drees & Sommer), Michael Leupold (Reos), Ulrica Schwarz and Michael Stöckigt (Strabag Real Estate), Tom Leppin (Reos), Frank Wellershoff (HafenCity University Hamburg).

Newly developed template to eliminate interface problems

Almost three-quarters of the eight work packages that comprise the research project, due to end this year, are complete. In addition, a template to address IoT gateways is already in place. The special feature is that IoT solutions can use this template to recognise and identify themselves to each other based on their functions, regardless of which protocol they use. This opens up the possibility of exchanging data between devices without having to disclose interfaces and program new ones. “We don’t want to develop our own standard, but rather to harmonise the existing ones,” says Reos project manager Michael Leupold.

Now that the cloud platform, which acts as a data hub, is to all intents and purposes ready for deployment, preparations are under way for the implementation of selected use cases. The tests being discussed are of networked IoT solutions for energy efficiency, user comfort and resource conservation. The choice of test property fell on the 20,000-square-metre Borx office property, which is currently under construction. It is scheduled for completion in late 2025 in the planned mixed-use Meltingport ensemble in Hamburg’s HafenCity. 

Only then will it become apparent whether the project gives rise to data standards that can be included in the VDI 2552 series of standards on Building Information Modelling. If that is the case, it will be a milestone in the development of sustainable properties – and resolve the problem of incompatible IoT solutions.

By Dagmar Hotze

Title image: Getty Images

More about these topics: