Will an AI chatbot solve passenger queries at Manchester Airport?

As the UK’s Manchester Airport began resuming its operations yesterday, after a two-hour power outage caused widespread disruption, there was one grain of hope that, in the future, passenger inquiries about cancelled flights and lost baggage will now have an extra platform for support – thanks to AI and the cloud.

Before the outage, TechInformed met with John Hudson, chief technology officer for Manchester Airports Group (MAG) at DTX Manchester, to find out how they are transforming the customer experience.

Even when things are running smoothly and even when travellers are experienced, a single flight can raise thousands of questions.

Where do I drop my bag? How long is the queue for security? Where can I fill my water bottle? Where’s my gate? Do I have time to get a coffee? How long has my flight been delayed? When do I arrive? Can I get a transfer?

For business or for leisure, the in-between of home and relaxing at a hotel is vital, and for many, stress-inducing.

Additionally, answers can’t always be found quickly on websites, the information desk is shut outside office hours – or the staff may not speak the passenger’s language. And a query on whether there’s a McDonald’s in the airport and whether it is before security or after it may seem too trivial.

For these reasons, Hudson, who joined Manchester Airports Group as chief technology officer (CTO) in 2022, claims that a friendly AI chatbot can provide the necessary assistance.

AI takeoff

As the UK’s biggest airport group, MAG serves sixty million passengers a year in its three airports: Manchester, London Stansted, and East Midlands.

According to Hudson, for 70% of people in the UK, one of its airports is a two-hour drive away.

“One of the things we realised is that we have a lot of information, but it’s not easily accessible,” he says.

“Some of it is flight information, queue times, or where water fountains, toilets, prayers rooms are.”

With this, it is building an AI chatbot that can be the front face of all its information and offer it to passengers on kiosks within the airport, or as a mobile application.

The conversational chatbot will be able to answer any question, whether it be flight time, where a toilet is, or if there’s a Starbucks in the vicinity.

Most importantly, Hudson says that the aim is to provide information more accessible.

Airports host many passengers of differing nationalities, so while the staff at the information desk may not be able to answer questions in a particular language, the AI chatbot will.

Another challenge the team is tackling is to ensure that the chatbot remains unbiased, ethical, and safe.

Hudson says that this means strict monitoring, testing, and ensuring that the correct policies are in place.

“As tech leaders, as engineers, as people in tech, we need to avoid the kind of issues that we’ve seen in the past,” Hudson says. “You need a diverse team; you need to understand all aspects of your application.

“Go out of your business and find people who are going to be using it. Diverse backgrounds, ages, genders, and sexualities. Give them your application and say, ‘Can you test it for me’,” Hudson advises other firms building AI applications.

“Guaranteed, if you have the same people who all look the same in the room testing something, it’s going to fail.”

For MAG’s chatbot, it was easy to tell it what not to talk about – politics, for example – but his team had to consider many other questions that could trip it up, such as ‘What can and can’t I take on a plane?’ and ‘Who should get given priority boarding access over someone else?’

“If people start asking these questions, we need to be able to answer them,” he adds – ensuring that the AI chatbot isn’t led astray.

Hudson also explains that the team is putting guardrails up to ensure it isn’t telling a customer their flight is late when it’s not, or a water fountain is on the opposite end of the airport to where it really is.

Partnership with AWS

The Manchester Airports Group chatbot is being built in partnership with cloud hyperscaler AWS. It follows the airport’s migration to the cloud.

“AI is one of those things where it’s so new, and I’ve got people in my team that want to build with it, but they don’t really know where to get started,” Hudson explains.

“What’s really good about the work with AWS is that its people come in and help the team understand what the problem you’re going to solve is, and what value the chatbot is aiming to bring.”

Hudson continues that AWS not only helps with understanding but also building, testing, deploying and then also standing by during the run and monitoring phase.

Alongside this, AWS is educating the team, “so they’re not just leaving a knowledge vacuum when they exit the project.”

Says Hudson: “I know by the time the application is up, and AWS leave, if I’ve got a new request for a new feature, the team I’ve got will be able to do it.”