Prologis and Home Depot leadership address the capabilities of AI for logistics

The ongoing emergence, as well as the potential, of AI (artificial intelligence) was a major theme of the annual Groundbreakers thought leadership forum hosted by San Francisco-based real estate investment trust company Prologis in San Diego this week.

In the opening keynote session, Ted Decker, Chair, President and CEO of Atlanta-based The Home Depot, explained that AI can be viewed as an evolution.

“On one hand, you can say ‘well, AI and Generative AI is overblown and just the next phase of an evolution of data capture,’” he said. “I don’t think that is the case.  That is a disproportionate and advanced spin on it…and certainly with the advent of the natural language models that power AI.”

Decker said The Home Depot has been on a 45-year journey regarding data, noting that in the early days it may have been a bit behind retail in general, whereas now it is on the cutting edge of retail. But in the early days, he described it as a matter of collecting data, getting data into appropriate data warehouse management systems, getting business management tools be able to mine that data and gain insight.

But now, as the company moves into learning models in computer vision and in how it is powering that in different areas, such as its web site search, which sees more than three billion visitors per year, he said it needs to understand their customers’ intent and also provide relevancy related to the search results.

“That is AI-driven,” he said. “Home Depot has scores now of PhDs in data science and as visionary as the founders were 45 years ago, they were probably unlikely to think that there would be PhDs in data science at the store support center here in Atlanta. So, we're using it for search. We're using it for customer care operations. We're using it to power interactions with customer queries. We're using it for marketing and marketing personalization and one-on-one contacts with e-mail in text. We're using it in our inventory forecasting and replenishment systems. The opportunities are endless in the power and the magnification of what you can do with that data now is truly at an inflection point that I think many of us are still very much in the exploration in testing phase. The deployment of these technologies over the next three, five, ten years are just going to be extraordinary.”

Prologis Co-founder, Chairman and CEO Hamid R. Moghadam said that industry stakeholders cannot be on the fence when it comes to AI deployment and implementation.

“[AI] is a big, fast-moving train, and you better get on it or you are going to get run over by it,” he said.

Moghadam observed that AI is not new, looking back at his engineering school days in 1977, when he took a course on it.  And he noted that ChatGPT and the large length language models took the idea of AI and applied it to retail.

“Now everybody knows about AI and that is what has really changed,” he said. “But before you can engage with AI and get value out of it, you have to have data and you have to have data digitalized or digitized. Without data, AI is useless. As companies have an advantage—companies like The Home Depot and Prologis—they have massive amounts of data, but most of them, especially in real estate companies, are not organized enough to digitize it. So that's the first step that we've gone through. And it takes a huge shift in culture because you got to explain to your people why you are collecting all of this information. Because you haven’t shown them how AI can help their business.”

The framework Moghadam said he uses to think about AI has three applications.

One is by using AI to do all the things Prologis is already doing but better and more efficiently, which he said “is interesting but not where the big prize is.”

The second application, he said, is using AI to make better decisions, for things like what to charge for a lease, how long is the lease it signs, and all decisions related to leasing with customers.

“We are using a lot of AI to drive those decisions,” he said.

He added that there is also a connection between AI and a focus on deploying massive amounts of capital ahead of its customers’ needs to make sure it is there for when a customer needs a facility.

“That means we need to figure out where to [put] some of these buildings,” he said. “That means we need to learn how to put EV chargers ahead of the needs in those facilities. So, capital deployment and where we invest our money is a very big opportunity for AI in our business. And the third application is a customer engagement tool. Everybody's going to be doing this. Out of our largest 25 customers that account for 20% of our business, there is a very long tail of smaller companies that don't have the sophistication and the resources to be able to do this stuff but they look to us to help them to do that.”

The Home Depot’s Decker followed that up, saying there is a need not only for data to effectively leverage AI but also to make it accessible through things like cloud computing, which he said is another big piece of what is powering AI efforts and does not require things like pinging a mainframe system or tying software to hardware.

“You get the data democratized in the cloud so all areas of the business can mine it with their various queries and build models,” he said. “For Home Depot, I think we have an efficiency and a virtuous cycle of productivity. The first thing is to do what we do better. An example of that would be inventory. For retailers, customer service starts with on-shelf availability, at least in the physical world or even in the e-commerce world. If you don’t have it in your DC, you can’t ship it to satisfy orders. AI is going to help us do what we always do, striving to get better on shelf availability and higher in-stocks. We've always been an anonymous mass retailer. What we are increasingly being known for is that we will be mass personalization.”