What airports can teach every facility about AI

In ABM’s recent What’s Possible Expert Panel series, Kayla Oliver, Head of Products, Partnerships, and Innovation, and Bob Clarke, who leads Client Experience and Operations Support, discussed AI and the future of facilities. In part one, we recapped their thoughts on why process, not technology, determines whether AI actually delivers.

Now, we turn to the industry where AI use has matured ahead of others: Aviation.

Ask facility leaders where AI is furthest along in practice, and aviation comes up almost every time. Airport operations are well versed in managing extreme variability. Unpredictable weather, shifting passenger volumes, tight turnaround windows, and zero tolerance for downtime make aviation an early proving ground for the shift currently spreading across every facility type.

That shift: from reactive, schedule-based operations to demand-based ones. AI now makes it possible to see, well in advance, that a particular week is likely to bring severe weather to a given city. Airports use that prediction to staff up ahead of the disruption rather than scrambling once it hits. The same logic now shapes decisions that used to run on fixed schedules and best guesses, from wheelchair deployment to restroom cleaning to how many staff show up for a shift.

Aviation adopted AI out of necessity. But any facility that still runs on fixed schedules, cleaning at set hours, staffing at set levels, maintenance on a set calendar, regardless of what is actually happening inside the building, is operating the way airports did before this shift. What aviation shows is less a set of specific tools to copy and more a preview of where every facility type is headed. Facilities are moving toward operations that respond to real conditions instead of assumed ones.

Three lenses for thinking about AI in a facility

Clarke frames the shift to demand-based operations around three ideas.

1. Is a decision predictive?
2. Does it draw on patterns?
3. Does it improve a process?

Predictive decisions anticipate what is about to happen, such as how many passengers are arriving on a delayed flight. Patterns recognize recurring behavior in how a building or terminal is actually used, not how it was designed to be used. Process means using that insight to change how work gets done.

Airports apply all three simultaneously. Flight data predicts passenger surges. Historical patterns show which restrooms see the heaviest traffic at which hours. And staffing processes adjust in response, in real time, rather than following a schedule set months in advance.

ABM Connect for Aviation applies that logic at the platform level, pulling in live flight and sensor data to anticipate passenger flow and route teams before a surge becomes a bottleneck, not after.

The pressure behind the shift

The shift in airport operations was triggered when the old way of working stopped working. Passenger volumes have returned to, and in many cases exceeded, pre-pandemic levels, while staffing has not kept pace. Traveler expectations have also risen, shaped by hotel-level hospitality standards, while regulatory and safety requirements have only grown more demanding.

“The aviation industry has recovered well, and the current priority is the passenger experience. Airlines and airports are passenger-centric, partnering to ensure that all their customers have a joyful, premium-feeling experience,” said Christopher Dohne, Vice President of Sales, Aviation.

Simply adding more staff or bringing more vendors to the already siloed third-party services landscape will not deliver those experiences. Airports that have moved to more coordinated, data-informed operations are treating the terminal as a single environment to manage, rather than a collection of separately scheduled tasks.

That coordination shows up in small, easy-to-miss ways that add up. Sensor data and heat maps now guide when a restroom actually needs servicing, replacing a fixed cleaning schedule with one that responds to real usage. This seemingly small change is a clear example of demand-based thinking in action: service triggered by what is happening in the space right now, not by what a calendar says should happen.

Charlotte Douglas International Airport shows how small changes can have a ripple effect. As passenger traffic climbed and the airport pursued its multi-year "Destination CLT" expansion, CLT's push to become an airport of the future required a service model built for scale and consistency, not one designed around a fixed set of assumptions about traffic and staffing. CLT partnered with ABM to implement technology-enabled oversight and dispatch coordination through ABM Connect^TM.

This solution and other ABM services helped CLT rank among the USA Today Top 10 Best Large Airports in 2025.

How to apply aviation’s AI lessons to your own facility

It would be easy for a facility leader outside aviation to conclude they are behind. Kayla Oliver says that’s not necessarily the case. "If you're in commercial real estate and hearing about everything that aviation is doing, that doesn't mean you need to do everything that aviation is doing," she said. "That's more using it as inspiration."

An airport's specific use cases, from flight-data integration, to wheelchair pre-positioning, to storm-driven staffing, are shaped by unique needs. A school, an office building, or a manufacturing plant will not need those exact applications. What they will need is the underlying discipline: paying attention to real usage patterns rather than assumed ones, and being willing to let a process respond to that data in something closer to real time.

That is a lower bar to clear than it might sound, and a far more useful starting point than trying to copy an airport's approach to AI. The question facilities leaders should be asking is, "Where in our own operation are we still scheduling around assumptions instead of responding to what is actually happening?”

In the next part of this series, we talk about the data needed to turn AI from promise to performance.

AI is only as reliable as the data behind it. What does it actually take to get a facility's data ready for AI? It starts with breaking down silos between systems and building strong governance.