Behind the Wing: How Ottawa's Airport Is Adopting AI | Ian Kerr

Andrew Faulkner, President, OPMMA 

At our latest OPMMA session, Ian Kerr, VP Operations at Ottawa International Airport, gave our community a practical look at how AI is being introduced in an environment where safety and security come first. There was very little hype in it, which made a pleasant change.

He started by asking whether anyone knew where the nearest first aid kit was. Nobody did. His point was that a safety culture depends on people paying attention to their surroundings, and systems alone won't get you there.

A changing airport

Passenger numbers are back to around five million a year, but the mix has shifted. Government travel fell significantly during the pandemic and hasn't returned, and Porter's growth is turning Ottawa into a connecting hub. The airport responded by building passenger personas covering business travellers, families, first-time flyers, people with accessibility or sensory needs, and northern communities travelling south for medical care. Each group needs something different from the terminal.

Applying AI, carefully

With revenue of about $160 million, the airport is a significant Canadian business, yet far too small to develop its own AI. Ian's team examines its own processes, identifies where AI could help, and asks vendors to show real value. A human stays in the loop on operational decisions, and nothing moves ahead until governance confirms the data foundation is in place and the risks are understood.

He used Gartner's AI Opportunity Radar to frame that risk. One axis runs from internal operations to external, customer-facing uses. The other runs from everyday AI, which Ian described as proven machine learning, to game-changing AI such as LLMs and agents. The result is four zones: back office, front office, core capabilities, and products and services. The airport started with low-risk internal uses such as predictive maintenance and is now moving into core operations, where AI supports decisions while a human keeps the final say.

The examples were refreshingly concrete. An autonomous baggage tug that initially stopped in snow and once braked for its own reflection in a puddle. Runway sweepers that calculate how much de-icing chemical to apply, with plans for driverless units in the middle of the convoy so both runways can stay open during storms. Video analytics that flag unsafe activity on the ramp. A gate allocation tool that makes recommendations while staff keep the final say. An app that converts gate announcements into text in the passenger's own language. It began as an accessibility tool and proved popular with almost everyone.

Trust came up repeatedly. Operators adopt new equipment once they've seen it work, so the airport runs new systems in parallel with existing practice and sends staff to see tools in use at other Canadian airports.

A familiar pattern

Ian's talk took me back to when firmware and algorithms first became part of our semiconductor products. Customers began expecting system-level solutions and wanted to see performance in their own applications before committing. We had developed a VDSL chipset, and our key customer, NEC, lacked the time and resources to build a modem to test it. To get NEC into field trials with NTT, our hardware team built a fully functional modem themselves.

Most organizations are in NEC's position with AI today. Few can build their own models, and they shouldn't have to. What they need from vendors, and from their own teams, is evidence that the technology works in their operation, delivered in a way that lets people build confidence in it. That is the work Ian described, and product managers have been doing it with every new technology for decades.

When asked how he keeps up, Ian said he hires people smarter than himself and has learned to synthesize what they tell him. He also offered a useful perspective on the current moment: AI feels seismic right now, but disruptive change has been a constant throughout his career. The 1980s, the internet era, mobile, cloud, and now AI. Each wave felt unprecedented at the time. His point wasn't to minimize what's happening today, but to reframe it. Change isn't the exception, it's the pattern. The organizations that do well are the ones that stop waiting for things to settle down and learn to operate well inside the disruption.