Smart machines

Everyone is talking about artificial intelligence. But what possibilities does it offer for the real estate industry? Quite a few – not least in property management and sales.

When the TV series “Person of Interest” debuted in the US in 2011, it was called science fiction. The story revolving around the ingenious billionaire Harold Finch, who created a machine to protect the population after the 9/11 terror attacks, seemed too far ahead of today’s technological possibilities: an artificial intelligence that identifies dangerous individuals and provides mass surveillance coverage far beyondhuman capabilities.

These days we can no longer be certain whether this artificial intelligence, simply referred to as The Machine in the show, is still a thing of the future. It looks very much like artificial intelligence – or AI – is in on the verge of a breakthrough. Amazon’s digital assistant, Alexa, has made her way into many homes; a Google program made headlines when it defeated the world’s best player of Go, an Asian board game. And a growing number of startups are developing business models based on artificial intelligence – including some for the real estate industry.

Intelligent analysis of contracts

This knowledge has not yet spread throughout the industry, though, at least according to a survey of more than 300 experts from different real estate sectors conducted in 2017 by ZIA, the German Property Federation, and consulting firm EY Real Estate. According to this study, the majority of those surveyed see potential in artificial intelligence, but don’t expect to see any impact until a decade from now at the earliest. Knowledge about AI is also limited: only 8 percent of professionals claim to have a solid understanding of the topic according to a study carried out by the Future Real Estate Institute on behalf of CBRE, a real estate consulting firm.

In principle, you can use AI for any process that has a simple structure and a high number of cases.
Gerald Kremer, Senior Manager EY Real Estate

For Gerald Kremer, this means market players are missing out on opportunities. “For me, artificial intelligence has enormous potential,” says the Senior Manager and digital expert at EY Real Estate. Kremer believes AI is not a futuristic fantasy, it’s already in use today – particularly in smart data and contract management as offered by companies like Architrave, Leverton and Evana. A survey by data room specialist Drooms showed how great the potential is for this technology, especially with regard to transactions. When asked which technological development will bring the greatest advantages for due diligence over the next ten years, 68 percent of real estate experts answered automated document analysis and 56 percent said smart data rooms.

“With artificial intelligence, data contained in digital documents can be extracted and analysed effectively, even in high volumes,” explains Jörn Stobbe, Management Board Member and Chief Operating Officer (COO) of Union Investment Real Estate GmbH. Not only does this simplify real estate sales, it also boosts efficiency in the management phase. “Where do an asset’s key indicators deviate from the market? What do tenants pay in similar locations? Would switching the energy supplier be worthwhile?” According to Stobbe, these are some of the questions that AI could answer.

To share in this potential, Union Investment acquired a minority stake in Architrave in 2017 (see interview below). This is not  the only collaboration of its kind. For example, investment company HIH Real Estate and auditing firm PwC are both working with Evana. “Our vision for this joint undertaking is the kind of AI that can take over commercial analysis functions in addition to capturing the data,” says Jan Brügelmann, Director at PwC Germany for Real Estate Transactions.

Building with a brain

Completed projects are already demonstrating that digital tools can increase a property’s efficiency throughout its entire lifecycle. For example, Austrian company PlanRadar has developed software that has proven helpful during the construction phase. This software captures and documents construction flaws with a web application and monitors any improvements needed. “To use it, you just walk through the construction site with a tablet, take pictures and write a short note – and the construction workers get started right away,” explains Constantin Köck, Technical Director and Partner at PlanRadar. This system can also be used to handle damage repairs in the operational phase. The results are impressive: according to PlanRadar, up to seven hours’ work a week can be saved.

“In principle, you can use AI for any process that has a simple structure and a high number of cases,” notes Gerald Kremer of EY Real Estate. This applies to automated bill processing, for example, and even more so to property management. “The possibilities for using artificial intelligence are huge in facility management,” says Klaus Dederichs, Associate Partner at Drees & Sommer, a consulting firm. One example is predictive maintenance: “A smart building knows itself when its individual components need servicing and schedules maintenance works accordingly,” explains Dederichs.

Smart solitaire: the Cube office building on Washingtonplatz is taking shape next to Berlin’s main train station and across from the government district. It is slated to open in 2019.
CA Immo (Simulation)


of the 2,500 largest publicly 
traded companies in the world 
in 2016 had a

But his vision goes even further. Dederichs is working toward a smart commercial property like the Cube office building currently under construction at Berlin’s main train station. “For that,” he says, “the building needs a brain that connects the building facilities with defined processes.” This brain recognises, for example, whether a meeting room has been used during the day. If not, the “brain” informs the cleaning crew that the room doesn’t need to be cleaned on that day.


In 2015, only 6 % had a 
Chief Digital Officer.

However, AI is also useful when working with end clients. Automated dialogue systems called chatbots can be deployed like human employees to communicate with renters and interested buyers via email or messenger service. For example, US-based company Apartment Ocean has developed an AI solution for real estate agents. Their motto: “Your personal real estate assistant that never sleeps.”

But why, despite this broad palette of applications and uses, are many real estate pros uncertain of what to make of artificial intelligence? Unclear terminology is one reason. “You have to distinguish between artificial intelligence and smart data management,” says Kai Panitzki, who invests in digital business models with his company, Bitstone Capital. “The current solutions focus primarily on data analysis. The next step would be to implement self-learning software.” Panitzki is referring to machine learning, which is at the heart of artificial intelligence. It enables the machine to recognise patterns and rules and then evaluate new, unfamiliar data.

This is exactly what PlanRadar is working on at the moment. The objective, according to company partner Constantin Köck, is to train the system to the point that it can analyse documents without human assistance. For example, it should be able to recognise that a photo is showing a hole in the wall and then automatically notify the relevant team that the hole needs repairing. The key is to train the artificial intelligence correctly. “Machines can also learn nonsense,” explains Thomas Herr, EMEA Head of Digital Innovation at CBRE real estate consulting firm. “The algorithms are only as good as the data that we make available. Incorrect data leads to incorrect conclusions.” Herr also wants to demystify artificial intelligence: “AI is a new set of tools, not a miraculous technology.”

Digital innovations wanted

For the second time, Union Investment and the German Tech Entrepreneurship Center (GTEC) are recognising the best digital solutions and concepts for the real estate industry with the international PropTech Innovation Award. The prize comes with a purse of €40,000 and will be presented in Berlin on 16 May 2018.

For the first time, the competition is being supported by the global industry association, the Royal Institution of Chartered Surveyors (RICS). And the award is also linked to the startup competition at the property market event MIPIM: the main prize winner of the PropTech Innovation Award is given a spot in one of the competition’s starting rounds and has the opportunity to present its concept at MIPIM in Cannes.

Information about the PropTech Innovation Award 2018:

New training courses are needed

However, real estate companies that want to use this toolset are facing a really practical problem: AI experts are rare and therefore expensive. In order to change this, EY expert Gerald Kremer wants to place greater emphasis on digital business models in property sector training. Progress is already being made in colleges and universities: in Germany, the Irebs International Real Estate Business School is running a two-week intensive course in Digital Real Estate Management in late February 2018, and in Switzerland, the University of Applied Sciences in Business Administration Zurich started offering a Certificate of Advanced Studies (CAS) in Digital Real Estate in 2017.

As long as experts are in short supply, established market players need to collaborate with technology companies. However, Gerald Kremer believes it’s important to place responsibility for digital transformation high up in the company – preferably with the CEO, who may receive support with operational implementation from a Chief Digital Officer (CDO). And the number of CDOs has recently seen substantial growth worldwide: according to an analysis by Strategy&, PwC’s global strategy consulting team, the number of companies with a CDO more than tripled from 6 to 19 percent between 2015 and 2016. The study looked at the 2,500 largest publicly traded companies in the world.

Data security and AI assistants

But companies also need to be vigilant when it comes to data protection and data security. It’s an issue that is “highly relevant for top management and board members,” emphasises Daniel Kautenburger-Behr, partner in the law firm Ebner Stolz. “The bigger the network, the greater the danger of security breaches by unauthorised persons accessing company systems filled with customer and employee data – and trade secrets.”

AI will also have an impact on the job market. Recently the influential Chinese technology investor Kai-Fu Lee raised eyebrows at an international conference when he predicted massive job cuts, particularly in the office administration sector. But he’s not the only one thinking along these lines. “Processes that are primarily completed manually today will primarily be handled by technology in the not-too-distant future,” says Daniel Kautenburger-Behr. And CBRE digital expert Thomas Herr is also convinced: “In 20 years, no one will even be able to imagine how we used to work without AI assistants.”

Maurice Grassau, CEO of Architrave, the Berlin-based data platform provider, is seen as a digital pioneer.

“We want to change the mindset”

Cooperation between established real estate companies and startups will help artificial intelligence achieve a breakthrough. A conversation with Maurice Grassau of Architrave

Mr Grassau, how do you explain what Architrave does to a property investor in two sentences?
We build digital assets, meaning we create an exact digital map of all relevant data and documents for each property. This includes an extensive spring cleaning of  the document repository.

What does this spring cleaning entail?
Our customers generally come with a laundry basket of documents that we throw into our machine. Then our machine tells us exactly what kind of documents they are. Based on the words occurring in the document, our algorithm recognises whether it is a land register record or a lease agreement, for example. It also filters out duplicates. This is all automated.

Classifying documents is one thing, analysing them is another. How far have you come with document analysis?
We’re talking about information extraction here. That’s the far more complex part of our work. Artificial intelligence is like a huge hammer still looking for the right nail. That’s why we need to analyse what information we need, and which algorithm is the best one for which document. We work with use cases here. For example,  we extract the relevant data from a lease agreement and check whether it matches the data in the customer’s SAP system. Or we review whether the maintenance frequency reflects the contractual obligations based on the maintenance logs. Human intelligence will still be needed to ensure the quality of the machine performing these tasks for a long time  to come.

Union Investment recently acquired a stake in Architrave. What do you hope to see from this collaboration?
It’s especially advantageous for information extraction. The key to a low error rate is having a high volume of clean training data which the machine uses to learn to recognise the rules. Generating this training data is highly complex and very resource-intensive work. Because we now have the lease agreements from all of Union Investment’s assets on the platform, we’re able to automate this work.

Why are you looking to work with established real estate companies?
I believe that digitisation will lead to cooperation rather than disruption in the real estate industry. We have data scientists, artificial intelligence experts and engineers – but training data is in short supply. Conversely, the major asset managers have huge data volumes from properties, but they lack expertise in artificial intelligence. So, working together offers a real opportunity for both sides.


What advantages does your solution offer when selling properties?
Incomplete documents lead to price reductions because certain risks are not covered. Smooth transactions start during the holding period in order to be able to realise the sale at any time. Our solution helps here because we can make it very easy for the asset manager to save the documents in the right location and identify which documents are missing or out of date.

Where could the Architrave solution lead in the future?
We do want to disrupt the real estate industry in one regard: we want to change the mindset that data and documents belong to the owner. We think they belong to the property itself. In the future, when a property is purchased, access rights to the digital asset will simply be transferred to the new owner, while the digital asset remains on the platform. Our vision is a single platform for all commercial real estate.

Media library

You can find video podcasts with more information about Architrave and PlanRadar in the media library at

Interview by Christian Hunziker.

As 2018 PropTech of the Year, Architrave is the winner of the Property Manager Award 2018. The jury praised the digital asset management platform for the quantum leap it has brought to asset management processes and for its considerable scalability. The platform’s users can trade assets and provide new owners with documents and data without losing time or quality.

Title image: Vincent Nguyen/laif, Colourbox

More about these topics: