Toggle Menu
En
You are currently on the Prologis global website
View Prologis Office Locations
Site search
Search Properties Near You
Global websites EN
Global websites You are currently on the Prologis global website
View Prologis Office Locations

GroundBreakers Magazine Logo Issue 9: Oct 2024

Tech Transformations in Logistics Technology: Insights from Prologis CTO Sineesh Keshav

How Prologis harnesses the power of its proprietary data across all business platforms.

Introduction

Asked how an aerospace engineer became the chief technology officer at a logistics real estate company, Sineesh Keshav emphasizes he never worked a day in his field of study: “I wanted to work for NASA, but as an immigrant I couldn’t get a security clearance,” he explains. However, NASA’s loss is Prologis’ gain. Over the past five years, Keshav—who previously oversaw technology at American Express, Safeway and Experian—has led a tech transformation to keep Prologis and its customers ahead of the curve, most recently by introducing the company’s proprietary AI platform.

Headshot of Sineesh Keshav

On his own time, he uses ChatGPT like a “personal vacation concierge,” including recently for a Fourth of July trip to New York City with his stepdaughter. “It was her first time, and she wanted to see the city. I said [to ChatGPT], ‘Give me a plan for each day.’ And it gave me a great itinerary, then rearranged everything to include the July Fourth fireworks, with commuting times from the hotel.” (See how other tech leaders use artificial intelligence to ease their everyday lives.)

We asked Keshav to share his thoughts on how technology is revolutionizing logistics at Prologis and beyond. Below are highlights from what he shared during our conversation. 

How do you transform technology at a legacy company?

I've been fortunate to work in some very different industries, but once you peel off the layers, the underlying tech problems are about the same: How do you manage a cloud migration? How do you provide an enhanced customer experience? How do you keep technology costs low but effectiveness high? When I came to Prologis five years ago, the company had its fair share of legacy processes and applications, like the fact that valuations ran primarily on spreadsheets. The IT organization was a lot more inward facing—making sure the trains ran on time, so to speak. We are now more outward looking in our use of technology, with dedicated technology teams to innovate in every business area. There's no one left behind on this journey.

We are now more outward looking in our use of technology, with dedicated technology teams to innovate in every business area. There’s no one left behind on this journey.
SHARE

What’s Prologis’ approach to hiring tech superstars?

Because our IT staff had been focused on looking inward, our technological transformation required a lot of new talent. We’re based in San Francisco, and it’s tough for us to compete with the likes of Google and Meta for technology talent. So even though this was way before the pandemic, we decided to be flexible and go to wherever the talent was—we didn’t expect them to move to San Francisco. Now, of course, everyone hires remotely. They were forced by the pandemic, but we were ahead of that and able to build a really good team with some of the best engineers in the industry.

Tech changes quickly. How do you keep up?

Because technology is moving so fast, experimentation and agility are essential. And that was initially a bit of a foreign concept for Prologis, given we are in real estate and there are no do-overs in real estate. (You don’t get to pour the foundation and then decide to start over.) But in software development, that's what people do; no product is ever completely finished. So, we embraced agility and started rolling out minimum viable products, a term that was not heard of five years ago at Prologis. Our people were not used to that kind of iterative work.

We realized our increased agility would require some change management, and in the past the Prologis IT team handled that, but it was not an ideal system. It’s like a restaurant where the chef is also running the dishes out to the tables. All the work in the kitchen stops. Plus, it’s a skill set mismatch because engineers are doing change management, which is not their strength. So now at Prologis, we let the chefs be chefs, and our dedicated Global Ops and Operational Excellence teams deal with training, adoption, KPI measurements and so on. Both of those teams have been a huge part of our data and digital transformation.

How do you design seamless workflows?

A key part of our tech journey is designing workflow. How do you design a day in the life of a remote execution manager in a way where they’re not pogo-sticking between apps, but there’s a natural rhythm to how they use the different apps to get the job done? We’ve made impressive strides. In turn, the new businesses we’re starting up have a chance to leapfrog on some of these transformations. For example, a dedicated team for our energy mobility and sustainability businesses is using our learnings on transforming technology in our core business. They’re ensuring the focus on data quality, business process automation and systems integration happens from day one. 

Where does AI come in?

AI is definitely going to be a game changer for us and have some underlying impacts on technology infrastructure overall. We now have PLDGPT, our own version of ChatGPT, implemented companywide. It enables us to integrate our data with the large language model, but Prologis is not in the business of building those large models. There are plenty of industry experts to build them for us, which allows us to be flexible and portable. Specifically, it promotes an application programming interface (API)-driven architecture where you plug and play the modules that surround your business, and any single part can be taken out and replaced.

When ChatGPT came along, it was clear there were no security protections for enterprise data. So, we worked with Microsoft to get access to the OpenAI large language model in a way that protects Prologis IP. And thus, we built PLDGPT, which we rolled it out in October, and we’ve already had something like 32,000 conversations. 

Beyond that lies generative AI, and we have exciting experiments underway. We see it as being both an efficiency driver, which is the most common use case for generative AI, and a potentially helpful capital allocation tool. We’re doing experiments on both fronts, and we'll come up with applications over the next few months.  We’re also building reusable frameworks so related used cases can be combined. An example of this would be the concept of PLDDocumentLLM, which is capable of assisting with document analysis and creation of all types of documents at PLD, whether it’s supplemental reports, Investment Committee memos or our contracts.

Graphic image of data

What’s your priority: how Prologis gathers data or the data itself?

Our tech experimentation goes beyond AI. One example is the Internet of Things, which feels like yesterday’s buzzword but is where a lot of work is still being done quietly. There are currently 80,000-plus sensors in Prologis buildings streaming real-time data into our data lake. That data is then analyzed for everything from helping fleet management for our mobility customers to helping bill our energy customers to alerting a maintenance tech that a pump room has a leak somewhere in their area of coverage.

All these different use cases have been enabled by this sort of technology. In this new world, the app is less important than the data. And we are laser focused on making sure we own our core data. The apps can be replaced—built or bought as technology evolves—but the data remains.

What does the future of tech hold for logistics?

End-to-end traceability and trackability of products in the supply chain has always been a challenge. It will require the individual components of the supply chain, Prologis being one, to be able to share data with the other components in an easy, secure way, where the consumer is the beneficiary of that visibility end to end. Different technologies, such as blockchain, have been tried, but the problem hasn’t been solved yet. I’m hopeful new technology will get us there. As more logistics companies embrace the cloud, and cloud-based technology infrastructure becomes the norm, we will end up in a place where collaboration between companies and industry alliances are essential.

Additionally, quantum computing is going to revolutionize how companies look at computer power. The new generative AI models use some very sophisticated GPUs and TPUs, so the need for computing power has gone up by orders of magnitude. Quantum computing may be one of those avenues where that innovation comes from, and that’s going to help companies like ours access sophisticated AI models at a very reasonable, ROI-driven price point.  

I would be remiss if I didn’t mention information security: The logistics industry will increasingly see cyber threats as a major concern that needs to be addressed. Recent incidents show how fragile the supply chain can be to disruptive events, so information security will increasingly become a focus for logistics companies to avoid such events.