Artificial future and solutions for travel

Edward Picture.jpg

Astute companies in the travel, hospitality and car rental industry should have, by now, realised the potential for Artificial Intelligence (A.I) to revolutionise their business. However, some firms still don’t know how it works, or who is using it and for what. In this piece CarTrawler looks at the ‘what, who and why’ of A.I, while also showcasing our cutting edge in house technology.

Society is engaging with A.I on a daily basis, these systems help with queries, map out journeys and even translate menus, all without the end user realising they’re being assisted. The goal of these systems is to imitate human intelligence processes by utilising machine learning and the large amount of data collected by companies. A.I can then predict customer behaviour and offer intelligent recommendations based on user preferences. According to Euromonitor Contributing Analysist Angelo Rossini “Automated and instant analysis of the huge amount of data companies collect about their customers will mean in the next few years an increasingly personalised shopping experience for travel consumers, which at the same time will result in higher rates of conversions into sales and in a more productive shopping experience for consumers.” [1]

Airlines like KLM and SAS have now begun to use A.I to engage with their customers through social media, as well as helping their staff to field customer queries. KLM recently revealed that their intelligent virtual assistants are now engaging with half of all queries[2] while SAS are building a surveillance bot that is designed to better detect fraud, especially money laundering. [3] Airports throughout the world have adopted A.I powered facial recognition and biometric technology systems, which assist human agents to securely process travellers at a much faster pace. Hotels and restaurants are also getting in on the action with A.I concierge and bar bots, that are currently deployed and actively dealing with customers. Meanwhile sophisticated AI systems are working away in the background of almost many new cars, helping drivers navigate and stay safe, while reducing accidents through advanced radar warning and braking systems. 

CarTrawler has developed its own in house A.I system called Arthur, which works with agents at the company’s Customer Centre of Excellence (CCE) to help put travellers at ease through every stage of their journey, by assisting with sales and bookings, as well as aftercare and support. Arthur is an innovative A.I auto response system that can analyse inbound email and web queries and come up with specific responses to a vast range of customer questions.

“CarTrawler is really beginning to see the benefits of A.I technology thanks to Arthur. In fact, we have obtained a 5X improvement in agent performance since we deployed the system. By doing the basics so well, Arthur has cut handling times for our agents from 210 seconds to just 42, which helps the CCE team to focus more of their attention on providing a personalised service for our customers, which is vital to CarTrawler’s continuing growth and success,” said Chief Customer Officer Michael Nolan.

Check out this episode of the CarTrawler Podcast, featuring CTO Bobby Healy. It’s about Arthur and why she’s great.  

Named as a play on ‘arthur-ficial intelligence’ the system is on track to revolutionise the way bookings are processed by the travel tech firm. Arthur uses natural language processing to understand messages received from real-world customers, the system can parse, cleanse, and extract meaning from inbound requests that are littered in spelling mistakes, incorrect grammar, slang words and abbreviations. Responses are generated based on the specific query and are tailored to individual customers before being handed over to human agents to approve or modify the responses.

Shane Lynn, who worked on the development of Arthur at CarTrawler, said that Arthur allows the company’s agents to handle simple queries faster. “Arthur takes simple tasks that agents are spending their time doing, like clicking copy and paste or responding to simple requests, out of the equation and allows the agents to get through a higher volume of requests.”

Shane was confident about A.I systems usefulness for any large customer focused company “any business that fields a large number of queries could benefit from a system like Arthur, the system’s efficiency is undeniable from a service agent point of view. However, there are also a number of softer benefits that Arthur brings to the table, like it never misspells words and will always ensure that best practises are followed, while also providing really good visibility for the administration staff on how the overall customer service department is running.”

There are two aspects to Arthur’s time saving abilities, according to Shane, “One is that it can pre-write and work out the best response for a particular issue and for those emails, we see an increase in process speed, because the agent can just review the inbound query, review Arthur's suggested response, and press go.” The second key part of Arthur’s time saving capabilities is the data it serves to agents once the email is processed by the system. “Arthur pulls all the relevant information and puts it in front of the agents. They don't need to switch time zones or open different software, all of the information from the query is in front of them, enabling them to review the question and Arthur’s response, and if needs be edit it.”

CarTrawler’s partner Ryanair, who are leaders in the travel technology space, currently have a bot in training on their production website. The airline are constantly deploying their Amazon Alexa based, voice controlled chat-bot, which they feed real customer interactions and responses in order to retrain and improve its accuracy, with the ultimate aim of reducing the number of calls going to a human agent. This voice-controlled chatbot will ultimately allow customers to speak into their smartphones in order to book flights, rent cars, arrange hotel rooms and more, on Ryanair’s app.

 “Everything on the website you can now do through the MyRyanair app. In the future chatbots and voice recognition will sit alongside the app to help our customer offering,” said Ryanair’s CMO Kenny Jacobs. [4]

At the recent CAPA Airline Leader Summit – CEO Gathering in Wicklow, Ireland, CityJet founder Pat Byrne spoke about how he values the importance of receiving a personalised service, how he wants staff to know his name and what he likes, and why some customer loyalty schemes don’t live up to this. This is an area where an A.I system could be invaluable to airlines. An A.I system could recognise that a loyalty customer is flying; it knows that she is a vegetarian so that meal option would be selected, they also know that she prefers a window seat in business class, which can then be offered to her by the bot. You could have a customer that flies to New York once a month, the A.I bot could automatically message this customer with options for the week he usually flies on, and ask which one he wants to book. Airlines could effectively employ an A.I system to act as a digital PA for valued customers. This is something that any carrier that values their loyalty scheme and business class customers should be investigating right now.

The Concierge in any top hotel is the link between the guest and the local area, advising them on every element of their trip, from restaurants and shopping to private transfers, tour guides and air transfers. Hilton Hotels have developed their own concierge bot called Connie. Connie is powered by IBM’s Watson A.I software and was developed to advise guests on where best to eat, visit and how to find their way around the hotel. According to Jim Holthouser, the Executive Vice President for Global Brands at Hilton, the company’s goal was to get rid of customer pain points, like waiting in line to ask a question and to help the hotel operate more efficiently.

The Director of Information Technology for Edwardian Hotels, Michael Mrini, told CarTrawler that their A.I concierge bot, Edward, was developed as a result of examining how guests wanted to interact with the hotel. “A large percentage of our guests were replying to our automated SMS system and we realised that some guests prefer to communicate with us through this method rather than picking up the phone, or going on the website to search for information,” he said. “We couldn’t possibly have a team of people just answering the text messages, so we decided to develop an automated system that can understand the question, and then find the answer and give it to the guest, or if it's a request for service, it can direct that request to the relevant department. That's when Edward was born.”

According to Michael, and TripAdvisor, Edward has been a hit with guests, “despite the fact that Edward introduces himself as ‘Welcome, I am Edward your virtual host’, many of our guests don't realise that it is a computer system. They think that Edward is an employee in the hotel. We can see it from their conversations with Edward and their feedback in the TripAdvisor comments, like ‘Edward was very kind and he sorted me out with late checkout’, in fact at one of our hotels, when we had just launched the bot, a guest left an envelope with some money in it for Edward. The guests love Edward as it has enabled them to get their questions answered very quickly, and without having to think ‘where do I get the information or who do I talk to at the hotel,” said Michael.

Michael told CarTrawler that he believes A.I systems like Edward will soon be the norm “I have no doubt that this is coming, not just the Radisson hotel group, all hotels eventually will have something like this because we have to listen to our customers, and our customers want alternative ways of communicating with us. They don’t want to be told to go on the website or to call the hotel. There is a whole new generation who want to communicate via chat. “We have seen, from the results of launching Edward two years ago that there is an absolute demand for this. So it's coming, definitely.”

“Guests and staff alike benefit from Edward,” says Michael, “It's extremely helpful for the customers and it's extremely helpful for the staff too. It's been able to enable employees, who would be sitting behind computers, checking emails, running reports, answering phones and giving out the same information again and again, to focus their attention on being hosts. “

However, A.I in the hospitality sector doesn’t stop at front of house departments as Hypergiant, a company that helps Fortune 500 companies to create machines that can learn, has developed an A.I system for TGI Friday’s. CarTrawler spoke to Ben Lamm, CEO of Hypergiant, the company responsible for TGI Friday’s A.I mixologist, Flanagan. According to Lamm, Flanagan “makes personalised cocktail and drink recommendations for the restaurant’s customers based on their interactions with the TGI Friday’s mobile app and loyalty program.” The system actually creates a flavour profile for customers based on a range of factors, leveraging conversational intelligence, machine learning algorithms and a reservoir of trade databanks.

“We wanted to create something that had direct engagement with the customer experience and provided value for patrons and the restaurant, while bridging the physical and digital customer experiences,” says Lamm. “What’s better than walking into your favourite restaurant and they already know what you like? It’s still too early to discern the business impact, but anecdotally customers and employees both love the capabilities. The magic is that it’s not really about the sophisticated A.I behind the system, but the interaction and experience for the customer.” The more Flanagan is used, the more it learns, and Lamm expects the system to “continue to expand its offerings, features and personalization. It’s all about customization and giving the customer what they want, with a little surprise and delight thrown in.”

Artificial intelligence isn’t limited to chatboats and virtual assistants, as A.I has been seamlessly making its way into the car rental market for many years through the motor industry’s integration of said systems into their range. A.I is used on a daily basis through voice commands, smart satellite navigation, crash prevention and autonomous drive systems. However, a certain level of Assisted Driving (AD) is currently available to the mass market; with self-parking cars in use on a daily basis that have 360-degree sensors, complimented by front and rear facing cameras, which allow the vehicle to position and park itself. These technologies have been available for a number of years, but have been the preserve of luxury, high-end vehicles, which appear in the premium section of most rental car fleets. However, Ford, keeping up with its historic reputation of bringing exclusive technology to the masses, has included this system in its new Fiesta. This car is armed with a night vision pedestrian detection system that is controlled by two cameras, three radars and twelve ultrasonic sensors that monitor 360 degrees around the vehicle, scanning the road ahead up to a distance of 130 metres. Utilising this hardware the car will deliver brake interventions to prevent bumps, will warn ahead of a possible collision, outline hidden pedestrians on the on-board computer and will also park for the driver. Many of these systems exist without the driver’s knowledge, but can be key in avoiding incidents, especially brake intervention and collision prevention systems.

People expect the utmost safety and security, which is why rental car suppliers should look more at publicising the underlying A.I safety systems many cars provide, as people who are uncomfortable about driving in a new country will surely take solace in the fact that the car will help them in an emergency. Furthermore, as A.I systems seamlessly enter our lives it is important that airlines and the hospitality industry offer their customers what they have come to expect in terms of accuracy and response times. Thanks to the digital age, customers now expect instant answers to their questions and immediate action on those queries. This is something your customer service agents will need some assistance with and that’s where A.I becomes a must, instead of a maybe.   

Morgan Flanagan Creagh