It has the potential to automate time-consuming administrative tasks, analyze vast amounts of data, forecast risks and opportunities, and enable workers in every facet of the commercial real estate industry to channel their time into thoughtful, impactful activities. But harnessing the power of Artificial Intelligence (AI) requires a prolonged, disciplined effort that encompasses every person and every operation in a company.
After six years of preparations, Patrick Hennessy, Director of Scheduling and Analytics at Harkins Builders, says the “fun is just starting. We are now at the very, very beginning of what we can do with AI.”
The company spent years assessing and standardizing all core digital processes.
Leaders of the effort consulted with Harkins’ 350 employee owners on how they completed each task and engaged them in an iterative process of testing promising software programs then refining and implementing standard practices. They migrated all of Harkins’ data — from business development, preconstruction, project management, financial management and other operations — to a single, cloud-based data warehouse.
And after extensive investigation, they acquired or began developing tools that could fully access and accurately utilize that wealth of data that was being rigorously and consistently fed to the warehouse.
The process taught Harkins a few things about how to conduct such a massive technology adoption effectively.
Broadly consulting employees was essential to fully understanding the processes and developing software-based solutions that would be effective, Hennessy said. It also became a way to unearth hidden strengths in the company.
“You would find these individuals that you never would have guessed it, but they had really intricate processes they were doing,” he said. “They didn’t seem like a big computer person. But when you gave them the opportunity to be involved in something like this, you realized they were the smartest person in the room. It was a goldmine.” Prologis also engaged in a multi-year initiative to create a world class, cloud-based, data warehouse to support its operations worldwide. Like Harkins, Prologis conducted sweeping efforts to involve all employees in data literacy, change management and technology adoption.
“We believe that every single employee, no matter what their role or where they are in the world, can use AI to improve what they do,” said Sineesh Keshav, Chief Technology Officer at Prologis.
Consequently, Prologis approached AI adoption in two ways. While company leaders worked on strategic, enterprise-wide technology decisions, Prologis also rolled out new digital tools to all employees to encourage “bottom-up innovation.”
“We have seen huge adoption,” Keshav said. “There is a natural affinity of people to see what this technology can do for them.”
For example, the company created a Prologis GPT — an AI tool like ChatGPT but with access to Prologis’s proprietary data — and made it available to all employees.
The result is “our employee population has built over 500 GPTs,” Keshav said. “Some are used at the individual level, some at the team level and some at the enterprise level.”
To be approved for broad usage, GPTs must be evaluated and demonstrate that they perform reliably and don’t hallucinate. And those new AI tools are already giving employees ways to complete certain tasks much more quickly while maintaining high quality and accuracy levels.
For example, Prologis employees created a GPT to customize letters to tenants. “We have standardized templates for these letters, but a lot of the information in the letter is personalized to the customer and it comes from their lease agreement,” Keshav said. Using the GPT, employees have “the ability to create a letter in a couple of minutes that would normally take 30 minutes because they would have to transpose about 15 fields of data from the lease into the letter. With generative AI, you feed it the lease and the template and train it to extract the right data elements.”
Similarly, Prologis employees have developed a GPT to rapidly create a customized, detailed one-pager ahead of a meeting with a tenant. The document can highlight any issues the tenant has had with their property, any interest they expressed in Prologis services (such as transitioning to onsite solar or adding EV chargers) and any known business challenges, such as a tenant’s desire to automate some distribution functions.
Harkins is also developing its own GPT to access and analyze the content of its data warehouse. That tool, Hennessy said, could complete highly valuable functions, such as looking at an upcoming project, comparing it to five similar past projects and then reporting on what went well or poorly with those projects, common challenges, financial issues, safety issues and so on.
“It could show me the risks I should look for in an upcoming project and even show me risks that I am not aware of,” Hennessy said.
That function and other AI tools could help Harkins operate more efficiently, better manage costs, avoid setbacks, deliver high quality projects and provide valuable information to clients, he said. “Ultimately, the data we collect on the projects we build for our clients becomes their data too, and that can help them with facility maintenance and property sales.”