What can airports and airlines expect from Agentic AI?

Airlines and airports have both been quick to embrace AI technology over recent years, but a more nascent development in the industry has been the implementation of Agentic AI.

Agentic AI is an AI system that consists of AI Agents, or machine learning models that mimic human decision-making to solve problems in real time, which can make decisions and adapt to evolving situations without the need for human intervention. Unlike the more traditional AI models, Agentic AI can exhibit autonomy.

Airport Industry Review sat down with Flyr’s chief product officer, Sam Chamberlain, to understand what the company has been working on with its partner airline, Riyadh Air.

Sam Chamberlain: What we’re trying to achieve is the ability to provide the disruption airlines need to be able to innovate and to change their digital retail experiences. For airlines and their passengers, there’s a big question mark now of where AI fits into all of that.

Flyr has been using AI in machine learning and making price predictions in our passenger revenue management products. That’s been around for a few years now, but we’re now looking at the shift to Agentic AI as that becomes more and more prevalent. So the rise of AI agents is happening thick and fast.

In the past, when making a price recommendation based on feeding an AI model, there was a lot of history, a lot of signals, a lot of information. That’s still a solid foundation for optimising both revenue for the airline and the passenger experience.

But Agentic AI is now stepping into that place of making a decision, implementing those decisions automatically, monitoring what’s going on, and being proactive. It's making a recommendation, but also implementing a recommendation.

What we're looking at is in the personalised offers space. When a passenger is shopping, when they’re looking for air travel and all the things around it, how does an AI agent make a smart recommendation of what is important to that passenger, based on everything that’s happening in real-time?

So, as you learn and as the agents learn about what the competition is offering, and what is important to this passenger, these AI agents are going to make more tailored and smart recommendations in real-time, responsive to the feedback a passenger is providing.

SC: Yeah, that’s a really good question, because historically airlines have – by virtue of the technology and constraints in place – really focused on passenger buying piece of the chain. So, a passenger comes and wants this airfare and books this seat, and that’s mostly it. They might change their mind at some point, or their mind might be changed for them if there’s a schedule change or a disruption. At some point, they get to the airport, they get on the flight, and that’s it.

But that’s a small piece of the journey. There’s the inspiration to travel before they even look to make a purchase. We call that inspirational shopping. How do you start planting the seeds with travellers that this might be the ideal destination for them?

Putting AI Agents into this type of space to make predictions on how we can get a passenger into the funnel, or how an airline can get a passenger into the funnel, is an opportunity.

The other thing is that airlines are now expanding beyond the confines of their own jurisdiction, if you like. So it’s not just about the flight, the WiFi on board, or the lounge access that you can offer the passenger – what happens on the way to the airport? What if the weather is bad? What happens if we know that it’s going to take a passenger longer in an Uber than it normally would to get to the airport? How do we make a proactive recommendation or suggestion that a passenger should plan to leave 30 minutes earlier, because of what’s happening in real-time right now?

These aren’t new concepts in terms of passenger behaviour. When you’re travelling, you’re checking your watch and you’re thinking about what you need to do to prepare. But it’s a new behaviour in terms of the airline taking some onus and responsibility, and helping you out with that. That’s where AI agents come into play.

They’re making these recommendations in advance of you even leaving your home, based on what is important to you. Is it spending time in the lounge? Is it getting through security as fast as possible? Is it having a nice view of the runway? Is it spending time in a retail environment? Is it having offers at a coffee shop? That then opens up this whole world of partnerships, and these are all things that the airline can now start to sell beyond their fare

SC: A couple of important parts to answer those questions. The airline ultimately has to decide by which method they are going to get in front of a passenger, which can change in different geographies, different markets, and different cultures. To give you an interesting statistic, in Saudi Arabia, around 90% of people book and arrange travel, and arrange everything around their travel on a mobile device. And so Riyadh Air is investing a lot in the digital experience, and that digital enablement allows them to get in front of a passenger better.

Mobile apps are a good example of that. SMS is still valid, but the second part of this, in terms of interacting with a passenger and knowing the passenger, is the shift to biometrics – especially when you’ve got the passenger in the airport environment.

Riyadh Air is very forward-thinking in terms of what the airport of the future and the passenger experience of the future will look like. They want this to be as frictionless as possible, so we provide the technology to ensure that there are no barriers. Riyadh Air doesn’t want a passenger to have to get out a boarding pass. In an ideal world, the airline doesn’t really want a passenger to have to get out a passport. Why? Why in this day and age do we still have passports? Why is that necessary? Our face is our passport, our biometrics, our currency.

SC: I think it’s a little bit preliminary. It's early stages at the moment. So we’re not having those conversations, but we would go hand in hand with an airline and prove that this is possible, and that the technology is there and the enablement is there.

So I do believe that at some point, a new Saudi Arabia is a good example, because it’s a massively growing hub and tourist opportunity. It’s going to be another great connecting hub in the Middle East.

They will build another airport at some point in the future, so now is the time for airlines like Riyadh Air to bake into that planning, in partnership with airport authorities. I strongly believe that with airlines like Riyadh Air in the fold and involved in the decision-making process, we will not see an airport like we have today. When that new airport is built in Riyadh, it will be the airport of the future. Now is the opportunity to say: "The tech is here."

We don’t need to do all these things the same way that we have done in the past. Yeah, security is security, but my mind races far ahead. So even at some point in the future, I think that will be very, very different. Technology will mean that your bag is being scanned as you walk through the airport, not when you put it on a belt and then through a machine, but maybe I’m getting a bit ahead of myself.

SC: If you travel today, if you travelled in the last few years, and your journey is disrupted, you will automatically have been getting notifications, maybe through the app or the airline, saying: "Sorry, your journey has been disrupted. We were late. You missed or you’re going to miss your connecting flight, and we’ve booked you on an alternative."

That, as a concept, is not new, but that’s achieved today through a very laborious framework of defining and hard-coding. Where Agentic AI comes into play is in more personalised recommendations. That AI agent knows that, okay, I’m going to miss my connection, but it’s still my preference to wait an extra 30 minutes and take airline A versus airline B. Or, my personal travel preference is, no matter what airline I’m on, I would always prefer a window seat over an aisle seat.

And this is all stuff in my head that I know I like. But if it’s in my head, it’s available for an airline to go and collect from me. When you sign up with an airline frequent flyer programme, or you log into an airline website, you can often store traveller preferences, for example.

Not all of this is taken into account today, but it's the type of things that agentic AI can take into account. I think that’s one way that it will make a difference, especially around things like disruption.

Furthermore, how are you capturing loyalty and bringing loyalty into the decision-making process through AI and data amalgamation? That is another opportunity. Think about every decision that the airline makes and that the traveller makes somewhere along this journey. Before they get to the airport, while they're travelling, when they're disrupted, after they get to their destination, and when they're at home thinking about their next trip. These are all things that are ripe for exploring and exploiting – to make them better, to give more choice, to give more freedoms, and to give more flexibility.

And I’m saying this like the airline has to do nothing now because the passenger has to do everything, because you’re putting it all in their hands. When I talk about flexibility and freedoms, part of that flexibility is the passenger's choice. Do they want to have the ability to decide those things for themselves and go in and action things themselves, or do they want AI or automation to go and do it for them? That is part of the freedom.

SC: It’s AI, and it’s a fast-moving space. It’s still maturing, and so it’s not going to get it right 100% of the time. The risks are that when you don’t get it quite right, are you losing the trust and the confidence of the passenger? Are you mitigating that if it’s not quite the right recommendation, if it doesn’t do quite the right thing, how quickly can it recover and regain confidence and trust from the traveller?

This is going to get better and more accurate over time. But the risk is that travellers also need to understand that it’s emerging, and there’s going to be times when it doesn’t get it quite right, but that it will get better and better over time. I think that’s probably one of the key risks.

To me, it’s about incremental steps and proof. You can’t say, look, we’ve got AI agents that will make automated decisions. So let’s go and automate everything under the sun, and let’s go and let AI decide every decision that needs to needs to be made. That’s a bit of a scary kind of thought.

I’m sure the volume of automated decisions, recommendations, and predictions in all spaces – travel and otherwise – will continue to increase. But the way to get it right is by taking small steps where you gain trust, confidence, and you prove that it’s doing the right thing. And, proving that you are able to pivot, quickly, if it’s not doing the right thing.