娇色导航

Worldport, the worldwide air hub for UPS, has made Louisville Muhammad Ali International Airport in Louisville, Kentucky, the third-busiest cargo airport in the US. The 5.2 million square feet facility boasts more than 20,000 employees, 580 aircraft (290 of them large-body UPS jets), and moves about 560,000 packages per hour.

“It’s a very intense operation,” says Alp Kayabasi, president of IT at UPS.

Until recently, asset tracking at Worldport involved labor-intensive, error-prone, and inefficient manual processes. Moreover, communication between load planners and ground crews relied on land mobile radio and was severely constrained. It required close proximity, from ramp to building, Kayabasi says, meaning decentralized planners had to be physically that close to facilitate any contact. Such a setup limited UPS’s visibility, hindered efficient operations, and prevented centralized management.

Identifying the challenges impeding UPS’s operations at Worldport was one thing, resolving them was another. But in June 2023, UPS started working on its solution called the Gateway Technology Automation Platform (GTAP), a way to leverage AI and machine decision-making to automate manual tasks, optimize resource utilization, and centralize exception management.

Kayabasi says GTAP helped UPS identify $13.5 million in savings in 2024, and has earned UPS a .

“This year, we’re projecting about $24 million in cost savings out of this initiative,” he says. “Beyond that, it improves our on-time performance, safety of our operations, and provides the best service to our customers.”

Weight of expectation

One area Kayabasi and his team identified early on was the need to digitize unit load devices (ULDs), containers to load freight and mail on wide-body and some narrow-body aircraft.

“We have over 60,000 of these ULDs all over the world,” Kayabasi says. “When an aircraft lands and is being emptied or loaded, there’s a swarm of employees around it doing everything from fueling the aircraft and maintaining it, to tracking the ULDs. The aircraft has to have a distributed weight and balance protocol as well.”

All these things, Kayabasi explains, were coordinated through radio communications.

“Having the devices electronically attached and our crew members efficiently communicating with each other is what we wanted to modernize,” he says.

The team created smart ULDs by digitizing them and integrating advanced sensors and communication practices to provide precise global tracking and real-time location services.

But designing the sensors for the smart ULDs wasn’t easy. There was no off-the-shelf, commercial solution that fit UPS’s needs. The sensors needed to be capable of providing frequent updates, and they needed to be shock resistant.

Battery management for the sensors turned out to be a sticking point, too. It would’ve been prohibitive to charge the device every time a ULD came off an aircraft. Furthermore, the devices were inside aircraft or enclosed buildings the majority of the time, so solar wasn’t a viable option. At the same time, the devices had to adhere to strict FAA and FCC regulations.

UPS engineered a power management system to ensure the longevity of the batteries without the need for frequent recharging. To meet regulatory requirements, the team worked closely with the respective bodies to certify the technology. They designed radio communication protocols that automatically adjust to a ULD’s environment to remain compliant.

Ramping up comms

In addition to the smart ULDs, UPS developed Ramp Chat as part of GTAP to eliminate Worldport’s reliance on land mobile radio. Ramp Chat is a communication platform to centralize load planning operations, and as a mobile application with multiple carrier capabilities and backup communication protocols, it ensures high degrees of reliability.

These two aspects of GTAP have given UPS the ability to leverage AI and ML to make operations even more efficient. An algorithm also determines which tugs — freight transfer vehicles — are best positioned to most efficiently load and unload ULDs, and AI helps balance weight and adjust equipment on aircraft, predictively determining where around the world additional ULD inventory is needed.

“The balance and movement of those assets are now artificial intelligence-driven,” Kayabasi says. “That was something we couldn’t have done without proper tracking of these assets.”

A collective effort

Kayabasi notes that a product management mindset and lean agile practices were key factors in successfully delivering GTAP by giving the team the ability to learn and deliver new features very quickly. His most significant advice, though, is to always use the business rather than technology as the lens to focus efforts.

“How do we ensure the investment we’re making in the technology moves the business towards their objectives?” he asks. “If their objective is better communication, how do you enable that? If it’s reducing their costs, how do you enable that?”

Above all, overcommunicate with your stakeholders.

“Change is hard, but by over communicating, creating training plans, and safety around the change you’re making, you can gain allies to help support you moving forward,” he says.

More on AI in the enterprise:

• 12 famous AI disasters
• Top 9 generative AI use cases for business
• Gen AI in 2025: Playtime is over, time to get practical
• Beyond the hype: 4 use cases that show what’s actually working with gen AI
• 10 ways AI can make IT more productive
• Rethinking and realigning IT for the AI era