Prologis combined its IoT-connected building management system with automation, and integrated both into the enterprise resource planning software that property managers use. “This platform encompasses reactive maintenance, equipment management, virtual document management, cost management and reporting, and data visualization all in one,” detailed Sineesh Keshav, the firm’s chief technology officer. Highlights include a pumproom management solution that is able to predict leaks before they happen, as well as a carbon data management system that helps further sustainability goals.
In contrast to its forward-looking sibling, generative AI, which uses data to actively synthesize information, has seen a more recent, yet explosive emergence into property management. The most immediate beneficiary is lease administration, which Ardern refers to as the “bugaboo of owner-operators.” Generative AI aggregates the data, but differs by way of its ability to generate the documents on its own. On the development end, Link created a proprietary platform that extracts the information from unorganized content, with the goal of training a language-processing model to function as a human would.
The advantage of such technology? “(It) gets smarter and learns every time it abstracts a new lease,” Elwell detailed. For Prologis, which uses AI-based document generation, “what used to take several days and weeks to get a lease into our operational system now takes several minutes,” Keshav observed. Of course, the personnel still play a role in auditing the abstractions and information.
For its part, Link Logistics plans to train a language-processing model to “interrogate the information in a human interaction kind of way,” one that allows customers and employees alike to “get data points faster, easier and smoother,” according to Ardern.
Other areas where generative AI generates value is in equipment operations, bringing complex manufacturing, HVAC and storage solutions online quicker. “Every piece of equipment comes with an information packet. The more we can leverage machine learning to quickly digest all of that information, the more we can plug it into the technology that our engineers are using, so that they have it top of mind, right at their fingertips,” Elwell detailed.
The ability to quickly and (most of the time) accurately create content also has vast marketing potential. Prologis is experimenting with using AI to generate marketing collateral. JLL has developed a proprietary GPT model called JLLGPT, that it uses from “writing great LinkedIn posts to leveraging abstractions of leases,” according to Elwell.
Down the road, Keshav foresees propensity to buy models and space planning simulations as “fascinating applications,” that need to be further refined to see widespread adoption.
In light of all these use cases, it may be tempting to adopt the latest GPT model or predictive analytics software, especially given the speed of their adoption by the market. “When you think about when the PC was introduced, it took 20 years to really impact productivity and operations. With mobile and cloud (technology) it was even shorter, and with AI it is going to be even shorter than that,” Ardern said.
But property managers would do well to consider the technology in the context of how humans themselves will use these models. “We can leverage AI to help us be smarter, but we can never want to replace the deep experience, skillset and knowledge of the people who are managing the equipment and these buildings,” Elwell cautioned.
Keshav agrees, and sees the time-saving aspects as the primary motivator. By saving our frontline employees’ time by automating repeatable tasks with technology and AI, we help them spend more time with our customers and other stakeholders,” he told CPE.
