‘There’s A Degree Of Wonder’: Why New AI Tools May Finally Be The Tech That Shakes Up Real Estate

Across the real estate industry, people are conducting their own tiny experiments to see how new artificial intelligence tools can help them do their jobs better.

Writing a first draft of a pitch. Finding a site for a retail tenant. Defining to whom they want to sell a product, service, or even a building.

Over the past two decades, the adoption of new tech has been a top-down process, and while real estate has been a laggard, AI tools like ChatGPT are being played with and prodded to see what they can do. 

Real estate may not be at the head of the pack when it comes to the use of AI. Yet large language model tools like ChatGPT, developed by OpenAI, are already being widely used, bringing the potential to dramatically change what it means to work in CRE, and raising philosophical and practical questions along the way.

“Everyone has heard of it, everybody wants to experience it,” JLL Technologies CEO, Leasing & Capital Markets Technology Amit Koren told Bisnow. 

While younger CRE workers are more likely to utilise AI technology on their own initiative, unprompted by the company, a huge number of staff at the giant global broker have been coming up with their own creative ways to use the new technology, Koren said. Many older staff might have heard about from their kids. 

“The degree of wonder associated with it is quite high,” Koren said. “Relative to any other tech in recent years, it is capturing people’s attention.”

The attention is so high that JLL is imminently set to roll out guidelines for its staff on the dos and don’ts of using new AI tools. One of the top things to avoid: entering confidential client information into one of these systems. Machine learning processes mean that this information may show up when others search for information about that client.

But Koren and his team will also be looking at dos, or ways in which JLL can more systematically incorporate the use of new AI tools into its business.

For those not quite sure what a large language model or GPT tool is, ChatGPT itself explains: “A GPT (Generative Pre-trained Transformer) tool is a powerful artificial intelligence model that uses deep learning to generate human-like text,” it said. “Trained on extensive amounts of data, it can understand and analyze natural language, generate coherent responses, provide insights, and assist in various tasks such as content creation, customer support, and data analysis. The tool's versatility and ability to generate text make it a valuable resource for businesses and individuals seeking AI-powered language assistance.”

The use of AI in real estate has been much discussed, but until the developments in the field of the last year or so, it has focused on more computational tasks like analysing large data sets. 

But that ability to ‘generate coherent responses, provide insights, and assist in various tasks such as content creation’ means it is more and more able to take on the kind of creative, insightful tasks it was previously assumed humans would almost always be better at. 

AI is still mainly being used to undertake tasks more quickly, Koren said, allowing individuals to be more efficient and do more deals. Firms that adopt this technology effectively will be have an advantage over those that don’t, he said.

Koren cited generating marketing materials or finding information on where the competitors of a retailer have sites as examples of how AI is being used. It can allow retail brokers to quickly collate data and use it to advise their clients on where to put a new store to exploit a lack of supply.

On the content-creation side, those interviewed by Bisnow were somewhat in agreement.

“It’s OK for a first draft, but it’s a long, long way off from replacing any of my team,” WiredScore Global Head of Product Jules Barker said.

One exercise his team undertakes is writing personas for different types of companies within the world of commercial real estate. As an experiment, the team got ChatGPT to rewrite some of those.

“It gave us some food for thought,” he said. “It wasn’t better than if a senior member of the team had written it, but it may have been better than a new grad.” 

In one of his famous essays, psychiatrist Sigmund Freud talked about ‘the uncanny,’ or how something that is nearly human, but not quite, can create a sense of repulsion in us greater than something that is obviously fake. For some in the real estate industry, AI-generated writing and communication evoke this uncanny feeling.

“I’ve tried ChatGPT a few times now to see whether it could help with the volume of newsletters and wider communications I write and share on a daily basis,” Yoo Capital Director of Communications and Social Impact Louise Page-Jennings said. Page-Jennings has been responsible for engaging with the local community in relation to Yoo Capital's £1.3B redevelopment of the Olympia conference centre in west London.  

“What surprised me most is, while the general messaging was there, the tone was always wrong. It felt unauthentic, formulaic, and almost heartless," she said.

"At Yoo Capital, we want to truly engage with the communities we invest in, and this should translate across all our communications. ChatGPT didn’t strike that tone for me, and to nurture relationships over written communications, genuine emotion and passion is key. I never say never, but for now I think the human touch is what’s needed in real estate, now more than ever.”

Proptech investment firm Pi Labs, on the other hand, has invested in one startup using AI technology, and it is looking at more. Contilio is a construction technology company that creates a computer model of what has been built on a building site in real time. If anything is wrong or out of place, it can quickly be rectified rather than waiting for a human to notice it. 

The firm is also considering investing in an unnamed startup that provides AI-powered conversational customer service assistance, via both text and voice.

“The real estate sector used to think of its occupiers as tenants, but now it knows it needs to think of them as customers, and provide better customer experience,” Pi Labs CEO Faisal Butt said. “Companies are trying to accelerate that change and level the playing field using tech.”

Tenants — customers — can engage with conversational AI tools to get a response to questions more quickly, as can potential clients who have made an enquiry about a product or building to a sales desk. AI means they won’t have to wait for a human to provide that first point of contact anymore.

Butt said the technology could even be used in the world of mortgage defaults, with an AI conversational tool finding out why a payment has been missed.

“The first few calls, while it gathers information, can take an empathetic tone, and it can be trained so that the longer the default continues, the tone becomes firmer,” he said. 

Yet to come, JLL’s Koren said, are the truly disruptive uses of rapidly advancing AI technology. That could include real-time rental pricing in all sectors, as increasing transparency means that instead of a human broker providing comps, rents in sectors like office and retail could be set using up-to-date data.

Such advances might be held back by the fact that particular data is hard to find in public sources, which is what tools like ChatGPT access to come up with their answers. That same factor is likely to hold back other possible advances in areas like the fight against climate change, WiredScore’s Barker said. Using sensors and analytics to make buildings more energy-efficient is a noble goal, he said, but no one has a portfolio big enough to come up with really insightful data.

If every office building in the world were sharing data, by contrast, it would be possible to come up with truly effective ways to cut carbon output. 

In terms of the future impact of AI, Barker pointed to areas like investment management, citing a recent experiment by analytics firm finder.com, which used ChatGPT to create a stock portfolio that outperformed the top 10-ranking UK investment funds over the past few months. 

“Investment managers, asset managers, leasing managers — I can imagine that your decisions are increasingly going to be supported by AI over the coming years,” he said.

“How resources are allocated is going to be supported by data, especially if these portfolios outperform over a longer period. You would question how many analysts there will be writing financial models when that can be done by AI. It feels like we’re on the cusp of something, and in a network-driven industry, building a business based on having privileged access to information no longer feels like such an advantage.”

That possibility raises the perennial question: Will the increased use of AI lead to job losses in CRE? Here’s what ChatGPT had to say:

"While large language models may lead to some changes in job roles and responsibilities, the commercial real estate industry is unlikely to experience significant job losses. The integration of these models presents an opportunity for professionals to adapt, upskill, and focus on higher-level activities, ultimately enhancing the industry's overall efficiency and effectiveness.”

Then it would say that, wouldn’t it?

In contrast, AI pioneer Stuart Russell, a professor of computer science at the University of California, Berkeley, outlined a different conclusion in a lecture for the BBC: AI will learn so quickly and so effectively that it will make all work obsolete, while concentrating unimaginable wealth in an even smaller number of hands. 

It is, of course, impossible to know who will be proved right. What is sure is that the real estate industry is already changing. Pi Labs’ Butt recounted how, during the interview process for a role at the company, he and his team noticed one question they asked received many similar answers. One candidate confirmed ChatGPT had written the response. 

Asked whether he thought this was cheating, Butt recalled the candidate’s answer.

“Of course I used it, if I didn’t, I’d be at a disadvantage,” the candidate said. “Would you rather have just me, or me augmented by ChatGPT?”

If you want to compare this article to a piece written by ChatGPT about AI and real estate, and decide whether Bisnow’s writer outperformed the robot, click here.