How to use AI and data to build operational excellence and resilience

Senior figures from across the built environment industry joined Yardi, a global leader in real estate software and technology, and BE News to discuss how AI and data are reshaping work patterns and adding strength to organisations.

INDUSTRY GUESTS:

• Pete Cooper, non-executive director, XPE Group
• Matt Glenny, UK and Ireland director of sales, Yardi
• Jill Ju, chief executive and founder, True North Management
• Akeel Malik, partner and fund manager, Urban Splash
• Sammy Pahal, managing director, UK PropTech Association
• Sanjeev Patel, managing director, PPP Capital
• John Powell, managing director, Frankham Risk Management Services
• Chloe Sachikonye, senior account executive, Yardi
• Steven Skinner, chief executive, HB Reavis
• Troy Tomasik, founder and chief executive, Outpost Management
• Will Turner, director – fund and corporate Services, JTC Group
• Liz Hamson, editor-in-chief, BE News (chair)

Increased use of agentic AI in the property sector could yield productivity improvements of more than 35%, but companies are being slow to adapt to the potential of the new technology, according to research undertaken by Yardi. The sector’s adoption issue was one of a number of topics discussed by a panel of senior figures representing companies operating across the built environment, at a roundtable debate held in Cannes, France, earlier this year.

The discussion kicked off with Yardi’s Matt Glenny presenting some of the key findings of the group’s latest study. One issue the study identifies is data fragmentation, which remains one of the main barriers to faster adoption of AI, with organisations keeping data in a variety of different systems and formats. As a result, Yardi has found that 88% of study respondents are piloting AI technology, but only 45% predict it will be the biggest disruptor over the next five years. Instead of being used for high-level strategic tasks, AI is often being used for low value, mundane jobs. “We’re at the very early stages of adoption in real estate, but other sectors are deploying this at scale already,” said Glenny. “The big barrier that businesses identify is data fragmentation; 86% say the biggest barrier to them realising these benefits is data being in lots of different solutions.”

According to John Powell, from Frankham Risk Management Services, that is an accurate assessment of the current situation. “Our industry is in a really bad place in terms of data fragmentation and data management. We often work with public sector clients, with local authorities, with social development providers. They have significant amounts of data within their organisations, but they cannot locate the information that they’ve got to analyse it,” he said.

Additional assessments

The situation is such that consultants are frequently called in for additional assessments because existing data cannot be found, he added.

“We’re at the very early stages of adoption in real estate, but other sectors are deploying this at scale already”
Matt Glenny, Yardi

“With the Building Safety Act coming into play, which requires property owners to make an assessment of their buildings holistically, AI has got a fundamental place in terms of analysing all of the datasets that currently exist within those organisations,” said Powell. This point struck a chord with Sanjeev Patel of PPP Capital, which has started to implement AI for customer support management processes.

“We haven’t yet figured out how we’re going to link the different data silos together, but we’re certainly implementing it for the operations side, and we have more things coming online over the next six months or so,” he said. Meanwhile, the sheer complexity of some businesses can make it hard to establish which AI packages and applications are going to offer the optimal solution.

As JTC Group’s Will Turner said: “Part of the issue of us working in so many different industries is finding tools that fit for all of those different scenarios. But also, we have a lot of software that might work really well.” JTC currently uses separate software solutions for the real estate and private equity sides of its business and another for employee share plans. “All that data is held in different places,” Turner added. “What do we need to invest in that creates a customer experience that is positive and that actually fits all the different platforms?”

“If we’re not answering a question within five minutes at 10pm on a Saturday, we lose them or they complain”
Troy Tomasik

According to Outpost Management’s Troy Tomasik, his company is currently implementing AI to enhance tenant experience. “It took a year to get the Q&A data warehouse together: the questions that you wanted to answer, the tone of voice you wanted it to give,” he explained. “But when it comes to residents and customers, if we’re not answering a question within five minutes at 10pm on a Saturday, we lose them or they complain.”

The built environment industry is already exhibiting an appetite for AI, and its potential Jonathan Vanstone-Walker uses and benefits, but many companies are still at the experimentation stage. Urban Splash’s Akeel Malik said: “There are two immediate ways we could be applying it: one is on the investment side, so looking at new acquisitions, and then the other is operationally, both for the actual customer experience of our residents and also reporting and understanding what’s going on within our assets.” However, there was broad recognition among roundtable participants that although many AI projects may seem relatively straightforward, they can be more demanding than initially anticipated.

TSP UK’s Jonathan Vanstone-Walker said his business has started using AI for customer service and the project has not exactly been plain sailing. “We were getting huge amounts of emails,” he explained. “So, we thought it would be one nice easy project to get a platform and be able to get people to self-serve. What we thought would take three months is now well over a year down the line, and it’s not done yet.” A lot of processes and information weren’t stored centrally, but “in people’s heads”, he added. “What was very good for us was that it forced us to write down a lot of processes and a lot of information about buildings.”

Prep work

For AI to be fully effective, a lot of preparatory work is required, agreed Sammy Pahal of the UK PropTech Association. “You first have to digitise those documents and then you can start to use AI to extract information from them” she said. “The UK government is recognising that.” This has resulted in a push by local authorities to get planning documents digitised more consistently, she added.

The scope of AI is huge, allowing companies to use it for a variety of tasks, big and small. But as yet there is no set playbook of how to use it, or how not to use it.

Yardi’s Chloe Sachikonye said: “I would say one of the most important roles that we come across is business change.” This process involves mapping out the systems and processes used, and required, by an organisation. It is a tough job to carry out while also running the business, and AI tools can be hugely beneficial in helping to establish a clean dataset and overall preparation, she added.

However, the ubiquity of AI can lead to security concerns, cautioned Sachikonye. “There are some operators that I speak to where operations managers and property managers use their own private ChatGPT to do work, and that is something that businesses should be alerted to,” she said.

According to Powell, such practice needs to be stopped immediately. Many new laptops come with AI assistants installed and this enables team members send out minutes of all video meetings instantly. “We tell them they’ve got to turn it off, because anyone can access it – we’re quite worried about that,” he said. After data fragmentation, concerns about security and confidentiality may be the next big barrier to wider AI adoption in the industry. Participants were not entirely convinced that confidential information shared with AI systems would truly remain confidential. Specific data, or insights from it, might then be accessed in answers given to rival companies, for example.

Instances of AI ‘hallucinations’ are also of concern. Another note of caution highlighted by some of the roundtable participants surrounded trust and the ability of AI to be 100% accurate. As Turner said: “We’re cautious about accuracy, because it’s all very well and good doing something quickly, but if it’s rubbish, it doesn’t work, that will lose confidence. I see hallucinations all the time, and when you ask, ‘where did you get that information?’ it’s just blatantly wrong.”

Experiences of such issues vary, and others are more trusting of the technology. “I’m not worried about hallucination, because that seems to be getting better, especially if you use the better, large language models. But the contamination of data is a real problem,” said Jill Ju of True North Management.

Verification stages

The more verification stages that an AI model goes through, the slower it will ultimately be. “Until we understand how to really do that, it can’t be that agentic, because then you are pushing out potentially wrong answers,” Ju added.

Powell said: “Confidentiality is a real barrier to the adoption of AI in the way that we like to embrace it. Dinosaurs like me who are at the end of their career don’t really understand it, we’re nervous about its adoption, so therefore you hold back on it. “Also, there is such a plethora of different systems that are coming into the market from software companies that are developing another model that could be useful. So how do I understand which models are appropriate to the efficiency of my business?” added Powell.

Despite concerns about data fragmentation and confidentiality, enthusiasm for the efficiencies that can be driven by AI remained high among the experts, the majority of whom were confident the technology would help them achieve their goals. “I don’t think it’s an option, really, anymore,” said Malik. “I think the industry as a whole will definitely get there [in terms of adopting AI].”

AI revolution

However, some companies may need to adjust their business models to truly benefit from the AI revolution, according to Yardi’s Glenny. “The operating model sometimes really needs to be changed in order for you to be able to benefit from it,” he said. “Some businesses will already have the data somewhere in some system that allows them to see, at a granular level, what’s happening with their tenants, what’s happening in their portfolio.” He added: “Many businesses we speak to don’t have that. They have financial information, high-level information that comes on a monthly or quarterly basis, and they have to pick up the phone to access that data. And so for those businesses, there is a big transformation to make.”

Ju said: “Some companies will get there. There will be a real bifurcation in terms of operational excellence, because it’s a challenge. The biggest challenge I see in the real estate industry is actually talent, because real estate really is quite a slow industry and the technology is moving so fast.”

The point about the real estate industry being slow to embrace change was echoed by XPE Group’s Pete Cooper, who cautioned: “I think this is the single biggest change to society we’re ever going to see in our lifetime. This is every science fiction film you’ve ever seen coming real right now, and it’s going to come at 1,000 miles an hour.

“I don’t think people have got any idea quite what’s going to happen and what impact it is going to have on the sector and on society.”

First published on BE News

---

AI is only as good as the data underneath it. Yardi connects your portfolio, properties, and teams into one source of truth – so AI can finally work for commercial real estate.

One platform. One dataset. Smarter decisions.