Analysis: AI in real estate - magical tool or limited aid?

Artificial intelligence (AI) is advancing at such speed that those real estate companies that are not already at least exploring its potential at this stage risk obsolescence within a matter of years. That was the key message to emerge from the PropTech Connect event held in London earlier in the autumn.

Dr. Kate Jarvis, co-founder of Fifth Dimension AI, perhaps put it most eloquently. ‘Have you considered the opportunity cost to your business if, five years from now, there is no one at any meaningful level of senior management who understands the potential of these tools and who has embedded them in your employees’ workflows?’ she asked. ‘You will be losing business to other companies. You will be haemorrhaging money.’

Of course, a technologist such as Jarvis would say that, which was a point she addressed directly. ‘Don't believe the hype - believe the money,’ she said. ‘There is a reason why companies and individuals have thrown an absolutely historic amount of money - billions and billions of dollars - at the development of large language models, towards the development of some sort of generalised artificial intelligence. We’re closer now to that technology than we have ever been.’

Jarvis invited the audience to compare the investment in AI to previous technological advances, pointing out that it is of a different order of magnitude entirely. ‘If you want to think of a contrast, you can think of past technical innovation waves, where the hype cycle around, let's say, crypto came and went,’ she said.

‘Blockchain as a technology has very narrow use cases. The same amount of money was never paid to it in the serious corporate sense that you're seeing today with AI. There's a ton of market share, a ton of capital that is going to be on the table to play for for those businesses that implement generalised AI data products first, best and fastest.’

It was a compelling case and one clearly being taken very seriously indeed by delegates. So, what opportunities does AI present to the real estate industry? How is it already being deployed? And what are the challenges to implementation?

Initial doubts and concerns
Of course, not all companies have embarked on their AI journey - or are still taking their first baby steps - and would be well advised to listen to those further down the road. Andrew McLennan, a managing director at Nuveen, says that a common mistake is to assume that AI can help with everything.

If such an approach is taken, businesses are likely to be disappointed or simply overwhelmed. Rather, he says, companies should identify specific tasks that potentially could be enhanced by the use of AI and then consider whether it is indeed the most appropriate solution. ‘The real focus should be on what the business problem is that you're trying to solve,’ he said.

‘AI is just another tool in the toolbox, in my opinion, and the input of course is data. ‘So, what the real focus should be on is do you have your hands on your data? And then if you do, can it solve a problem? But then also focus on, what you are trying to do with the data. AI might be the right solution. But it might not be - there may be classical solutions.’

By starting small with well-defined tasks, businesses can also avoid the risk of spending a potentially huge amount of money and seeing little in return. ‘AI is not a magical tool,’ said McLennon.

‘It won't solve all of your problems. It's a tool to solve specific problems. So define your problem. Define the investment. If you're spending a million dollars, for example, on some AI solution, that's all good and well, but if it only generates $10,000 of benefit, you're just adding costs without getting the actual return.’

However, McLennan added that while the application of AI may currently be limited, it will improve rapidly and will soon move from fulfilling clearly defined tasks to providing potentially strategic advice to senior management. ‘People always say that AI is not that good yet and I always equate AI currently to being like an intern in your team - it’s overly enthusiastic, but doesn't always come up with the right answer,’ he said.

‘But as we give it more data, give it more computational power, and as we become more intelligent about how we use AI to solve problems, it will go from being an intern to being a junior to a senior and eventually a partner equal to yourself. You will have someone sat next to you in your computer who can think like you and can help not just answer questions for you but give you ideas about how you should be thinking and provide insights.’

Current uses in real estate
One of the principal ways in which AI is currently being used is to rationalise data contained in a huge number and various types of documents and then use that data to make useful observations. ‘A lease is never just one document; it’s a huge pack of documents,’ said Jarvis. ‘There are legal documents, there is handwriting. It's a huge pain to go through that manually and even if you have something to extract those data today, they're not very accurate. Someone is still sitting there checking.’

However, Jarvis noted that the latest generation of AI tools aimed at the property industry is changing all that. ‘What a product like ours does is extract that data with increasingly high accuracy, so 98% or above,’ she said. ‘It’s then about serving it to you in a way that's intuitive and easy to use.’

David Richards, director of technical infrastructure at Google, said that he is focused on a quite different area: how AI can be used to make buildings operate more efficiently - and consequently in a more environmentally sustainable way. ‘I'm particularly passionate about where we can use AI to make our buildings just work better or, in some cases, just make them work,’ he said.

‘It’s absolutely about improving the experience of people; people living and working in those buildings. We're quite focused on starting to pilot some of the things that will really help make that work. We do a lot of work on making our buildings and our spaces frictionless, so you can move through your working day without technology being a problem. It's just a slick, whether it be the way we book meeting rooms or the way we release meeting rooms when they're not being used.’

The approach involves aggregating and then using AI to interrogate data from a vast array of sensors in order to understand how a building is being used. ‘For instance, there is the idea of using temperature feedback data to work out how occupancy is working and then relax set points in buildings so that we use less energy when bits of them are not occupied,’ Richards said.

‘We have a really fascinating piece of work around refrigerant leakage, where we're trying to stop leakage of refrigerant happening as early as possible by just measuring pressures and temperatures in the refrigerant loop. That means we make those systems work more efficiently and they leak less refrigerant, which is good for the climate.’

In fact, Richards added, Google is already taking the technology to the next stage by using AI to not just diagnose problems but to come up with whole building solutions. ‘We're doing a lot of work with diagnosis, but then linking that to large language models so that we can actually work out where our buildings aren't working optimally and have an automatic proposal of what to do about it,’ he said.

‘By doing that we make the process of operating a building much more efficient and, again, drive carbon use and water use as low as we can. I think all of those things are experiential. Some of them are about the experience of people who actually have to be in our buildings, but [there are] other benefits.’

Another company that is already taking advantage of AI is Matterport, which, among other things, produces 3D cameras and a software platform to help clients automatically create 3D tours. ‘We’re seeing a lot of adoption and use in the field of property marketing, where people are using AI to generate all types of assets to help users engage with a space, whether it is already built or new construction,’ said Mike Montgomery, a vice president at Matterport.

‘The input can be very simple. It can be as simple as a picture and then using AI we can derive all kinds of information from that simple image or those images. And then it gets exciting, because we can layer additional data on top of that spatial data. That's really where I see it as a stepping stone to unlock a lot of potential.’

Data privacy and cybersecurity
The potential for hugely positive change isn’t in doubt. Challenges remain, however, not least when it comes to data privacy and cybersecurity. ‘There is a huge amount of seriousness with which we treat data security and data privacy, and the amount of air time we give it is very well justified,’ said Jarvis.

‘These datasets are our livelihoods. They're also people's lives. There are tons of personally identifiable information that's contained in transaction documents and in built world environment data that shouldn't be leaked to the general public and could easily be used for malice.’

Ian Boggis, head of data strategy and operations, data and insights, at Cushman & Wakefield, agrees that companies need to be keenly aware of their own data security as well as that of their clients. ‘From a regulation and compliance perspective, we seek to ensure, whether we're working with a third party vendor or leveraging our internal AI development hub, that we have security processes in place,’ he said.

‘Every new solution is reviewed and taken through a review board, which also consists of data privacy leads and compliance leads. So, we look at every solution from a security and data privacy perspective when it comes to adoption.’

Richards points out that Google takes a similarly robust approach when it comes to building operations, not least when it comes to procuring companies that work in its supply chain. ‘Zooming in on the operational side of our buildings, we consider the vendors we use very, very carefully,’ he said.

‘We put them through very rigorous security protocols. Even if the data isn't actually leaving our ecosystem, we're still very careful. And if the data is leaving our ecosystem, that process is even more rigorous. So there's a lot of rigour around the vendors we use and how they capture data. I think it goes a long way beyond legislation.’

Moreover, Richards added, Google is similarly robust when it comes to the connected devices used within its buildings. ‘We all need to get to a point where the data that constantly flows from all the devices in our buildings is also secure, so we are similarly rigorous about putting any device that goes into our buildings through a rigorous testing regime around security,’ he said. “[That’s about] a device in its own right and a device connected to systems. We need to make sure it is secure.’

However, Jarvis observed that while companies do indeed need to take security seriously, it should not be an impediment to AI adoption. ‘From a Silicon Valley technologist, I have to say that these challenges are not new,’ she said. ‘These are the same challenges that people have been facing in building technology for decades and decades. They're just perhaps amplified by the amount of data and the number of ways in which that data can be breached.’

So long as companies have the right intentions and work with intelligent partners, Jarvis added, there is no reason to be paranoid about security lapses. ‘While there are some concerns that exist, they exist because companies don't care to be compliant, not because they technically can’t,’ she said.

‘Let's be very clear, there are easy engineering solutions. It’s very possible with the right amount of time, resources, money and the right architecture, the right engineering, to create data compliance solutions. [There are some businesses] that want your data to use for their own profits, but there are also companies that are there to really service an enterprise’s data security needs. You just need to be able to differentiate between those two things.’

Engaging stakeholders in the process
Then there is the small matter of take-up. While some employees may well be AI enthusiasts, others will be resistant - and property, it has to be said, can be quite stuck in its ways. Taking people with you is therefore essential. At Google, for instance, senior management realised this some time ago and, as a result, all members of staff receive AI training no matter their role so that they can see the technology’s potential.

However, you don’t have to work for a tech giant to realise that working with AI demands a change in mindset. ‘I've been doing this for 40 years in the technology space and it's always about the change management piece,’ said Roger Wahl, chief technology officer at Innovo Group. ‘Technology is usually not the issue - it’s willingness.

‘It's taking the business stakeholders on that journey with you, instead of forcing this on them, to eliminate the resistance. So, when you come to propagating AI through the culture of a company, this is the challenge that you'll have. It can work in pockets and you can train super users on the generative AI side, but it’s going to take change management and nurturing to make this part of the corporate fibre of what you do.’

Montgomery agrees. Indeed, he believes that it is only by boosting familiarity and engagement with AI across the entire real estate industry that the sector will truly take a quantum leap forwards. ‘I think the more we can enable and make it easy to engage with and consume that technology, the more it's going to be used by businesses,’ he said.

‘That way, more real estate companies are going to say “okay, we can do this”. We can see how that can help us change the way we do something for the better, whether that's an internal benefit where we improve our efficiency, or an external benefit where we engage with our customers in new and better ways that meet them where they want to meet us.’

Big picture
So, deploying AI to best effect requires both attention to detail and strategic thinking, but overall the property industry is waking up to the opportunities. Moreover, nobody should forget to consider the bigger picture, according to Dr Mahmoud Al Burai, senior advisor at the Dubai Land Department.

‘We're looking at the whole ecosystem, the whole life cycle of real estate,’ he said. ‘At the end of the day, our vision is that we want to build a better city; a more sustainable, resilient, affordable and on top of that a safe and tolerant city.’