Outside of the airlines industry everyone talks at the moment about artificial intelligence, specifically generative AI. From where you sit, how is that transforming your industry?
That this was the question that kicked off the APEX EXPO Thought Leadership sessions in Long Beach last month, setting the tone for multiple interviews throughout the week, says a lot about the industry. How executives answered variations of that query was especially telling. Particularly given the different impressions they have of how Generative AI works and what it is capable of delivering today.
Virtual Agents for Support
That initial AI question was posed to Air Canada CEO Michael Rousseau. He responded with optimism, sharing “there’s no doubt that we can leverage it as an industry and certainly as individual airlines.” The carrier now has 75 full time employees working on AI endeavors, plus a new Chief Digital Officer, overseeing improvements in data services and AI integration at the airline.
Much of that work is going into productivity efforts, “a lot of different applications around OTP, schedule optimization, maintenance optimization.” Rousseau is keen on those, particularly given the cost savings they promise. But, he continued, “What excites us more is on the customer service aspect. Air Canada has one of the best concierge services in the world for our top customers. We would love to offer that service to all customers. We think an AI concierge agent, a virtual agent, could possibly do that for all our customers.”
Air Canada is not alone in pursuing this use case. Anton Vidgen, Riyadh Air‘s VP Guest Experience, offered a similar take in a separate session. Thanks to AI and passengers’ willingness to share their personal data, Riyadh Air expects to provide “instant access to key guest information for frontline staff.” Ultimately that converts to “hyper-personalization, the idea of one-to-one at scale.” The primary example used for this hyper-personalization is one others at the airline have frequently shared: cabin crew will set the tray differently for left-handed passengers. It is unclear how compelling that is as a passenger experience, but it is what is public so far.
A quality concierge program is a high-touch experience. Several airlines offer similar services for premium cabin travel, and that dedicated resource throughout the airport experience is often well received. And Vidgen acknowledges that the human touch is key for the truly top-tier experience. But the virtual agent is one of 200 use cases the carrier is pursuing.
Unrelenting Speed
Oman Air is in the middle of a restructuring as it, like many others, works to shore up its finances. To that end some of the statistical and machine learning aspects dubbed “AI” are helping the carrier to find efficiencies in its operations to trim costs. Those efforts are successful so far, according to CEO Con Korfiatis. But the speed at which the digital landscape is changing is challenging for the company.
“I can see the benefits of the small steps we’ve made so far and just how quickly it leads to transformation. And that’s where the world is headed,” he explained. But it also means not having sufficient time to analyze the options and possibilities. Korfiatis continued, “We don’t have time to sit there, so we’ll study something for a year or two years. Now you’ve got to say, how do we apply some technology to this? And how can we have an answer for tomorrow morning when we can look at something and start making decisions?”
Korfiatis also shared an anecdote, attributed to another airline CEO who asked the IT department how many different systems were in service at the carrier. The team came back with a spreadsheet detailing 155 technology platforms in use.
“He looked at that and thought, I don’t know how to deal with it. So he dropped the spreadsheet into chat GPT and said, ‘I’m an airline, and this is what our IT infrastructure looks like, and this is supposed to deliver talks to each other. Can you tell me how to simplify this? How many do I really need? And how many can I do without?’ And it came back saying the airline needed eight out of the 155.”
Korfiatis lauded the near immediate speed with which those results were returned, particularly compared to the time it took for the initial list to be developed. But he also appeared to gloss over the fact that the answer given by ChatGPT was almost certainly wrong.
A Dirty Data Problem?
Rousseau’s optimism on the value of Generative AI remains unimpeded by the black eye Air Canada took earlier this year fighting a customer who it says was misinformed by the chatbot on its website. Ultimately the carrier lost the lawsuit, despite arguing it was not responsible for that content since an outside company operated the bot. Rousseau suggests it was a one-off issue, overplayed in significance by social media. Moreover, he said the underlying problem was bad data in the bot’s source database, not an mistake or hallucination made by the technology. And he’s willing to take some of those mistakes for the sake of what he sees as progress:
Will mistakes happen? Probably. But, you know, we’ll get better. And GenAI will get better at solving those things. We have good governance models now to ensure the data is clean and we test a little more rigorously. But mistakes are made. I’ll live with that, as long as it’s not too damaging, obviously. Because I know the greater good will help a lot more customers.
When pressed to commit that he would not fight back against future customers who were wronged by the carrier’s AI systems, as it did in the chatbot case, Rousseau declined.
Sizing up the Competition
With its “digital first” mindset and 200+ AI use cases on the roadmap Riyadh Air has plenty of other opportunities to put those systems to the test. Vidgen shared that the company has a market sentiment solution scanning social media and other sites for details on what passengers are saying about a broad range of carriers that Riyadh Air expects to both compete with and cooperate with as it grows. “I’ve got a dashboard that tells me what do people think about the food on Delta Airlines,” Vidgen shared. “And as we start thinking about different markets we’re going to be flying to, we can drill down and develop appropriate products and services to make sure that we’re going to be able to competitive.”
He also shared that the carrier is using a 3D body scanning tool “to get much more accurate measurements for our cabin crew and over other front line employees.” Given that the clothes are not custom fit it is unclear how that’s better than having the crew simply try things on. But it does allow the carrier to push the “digital mindset” concept a bit further.
Airlines have not been shy about trying to position themselves as leaders in the digital world, successful commerce companies that happen to fly airplanes. Success on that front, however, has been incredibly limited. Many at least looked at blockchain and NFTs, though few truly invested along that path. Far more appear committed to Generative AI, given the promise of better customer service with fewer human resources required. And they seem to be OK overlooking the part where it doesn’t quite deliver and isn’t really getting better, despite promises it will and billions of dollars invested to make sure it does.
A favor to ask while you're here...
Did you enjoy the content? Or learn something useful? Or generally just think this is the type of story you'd like to see more of? Consider supporting the site through a donation (any amount helps). It helps keep me independent and avoiding the credit card schlock.
Leave a Reply