Airlines put AI on the frontline of customer service

It's not just technology companies that are putting artificial intelligence (AI) first these days. AI is becoming the new self-service and airlines are among the first big businesses to incorporate these new technologies into their customer experience offerings.

A few years ago Heather Pemberton Levy, vice president of content marketing at Gartner, predicted that 'by 2020, the average person will have more conversations a day with bots than they do with their spouse'. Married couples may have to wait until next year to find out whether this is true, but the advance of conversational AI, both chat and speech-to-speech, now continues apace.

According to a global survey published earlier this year by software company Salesforce, 53 percent of service organisations said that they expected to use chatbots within the next 18 months.

The airline business, which handled an estimated 4.4 billion passengers during 2018, deals with some of the highest volumes of incoming customer enquiries of any industry. As a consequence, peak travel booking periods can lead to long call wait-times, lack of availability of call centre agents and unresolved cases. This is one of the reasons that airlines are incorporating live chat into their customer service offerings. Worst case: if it takes a while to get a customer service agent on chat, at least you haven't been left hanging on the phone for a few minutes. However, with new AI chat platforms, an airline's customer service response can be instantaneous.

Asian airlines are among the first to embrace this new customer service trend and many airlines across the continent are already using artificial intelligent agents to answer customer queries, provide flight information and even book flights.

In August, Malaysian low-cost airline AirAsia announced that AVA, its AI-powered chatbot introduced in December, has contributed to a significant decrease in the number of customer queries handled by its call centres. This is becoming a common pattern worldwide and, according to Gartner, companies are reporting reductions in customer calls, email and other enquiries of up to 70 percent after implementing this kind of technology.

In fact, AirAsia seems to be phasing out customer service calls, closing 8 out of its 10 customer call centre operations. Of course, the ongoing effort to keep operating costs down is a key factor here, but there are many customer experience and revenue drivers behind this AI-first strategy too. AirAsia is investing in artificial intelligence technologies to improve its customers' travel experiences and it credits AVA with increasing customer satisfaction by 90 per cent over the last 12 months. The chatbot has also reduced wait-time in the airline's response to customer queries.

Ruth Zive head of marketing for the Toronto-based customer engagement platform Ada Support that developed AirAsia's chatbot, explains: "An automation-first strategy frees live resources to engage in more mission-critical, high-churn interactions, which ensures a more meaningful customer experience."

"A chatbot like Ada accommodates an airline customer on the channel of their choice, 24/7 - to upgrade their ticket, change their meal, track their flight, or change their reservation - all without the need for live support."

So, why is it that almost everyone who has experienced chat support has had at least one unfulfilled, time-wasting encounter? This is still common, because many chatbots are still using old technology, while others are poorly thought-through. AI chatbots have been around for many years, just not in a form that are easy to develop and deploy at scale for businesses.

Frustrated with poor functionality, high costs and user dissatisfaction, most large businesses haven't made chatbots a priority until the past five years. New technology developments, new machine learning algorithms and more user-friendly development interfaces have helped encourage adoption, but a major force driving chatbot development has been the soaring volume of user conversation data available for AI and its programmers to learn and make rapid improvements from.

Apple's Siri launch in 2010 opened the world's eyes (and ears) to the possibility of natural language processing (NLP), but it took Amazon's Alexa launch in 2014 to begin to convince business that NLP was nearly ready for prime time. NLP is a branch of artificial intelligence that allows people to use natural language to interact with computers. It's allowing more and more natural conversation between devices and humans all the time, which can support an ever-increasing variety of services: from shopping, to entertainment, healthcare, or education.

AI chatbots are currently being adopted much faster by big business than AI voice assistants due to a difference in development costs, availability of enterprise- ready platforms and ease of deployment. According to professional services firm Deloitte, we're now in the midst of the third phase of the chatbot revolution.

Early scripted chatbots looked for key phrases and gave a limited number of pre-defined responses, while more advanced versions were better equipped to identify a customer's intent and fuller meanings of user conversations.

Now, chatbot development has now entered its third stage: the realm of the virtual assistant, where developers are producing chatbots with the ability to have end-to-end conversations and provide solutions to customers drawn from other intelligent systems. The fourth stage is the holy grail of artificial intelligence, the human-like advisor that has the capacity to reason, but we're not going to meet this AI anytime soon.

AirAsia is not the first Asian airline to introduce an AI chatbot for customer service. Japanese carrier JAL launched an AI-powered assistant called 'Makana-chan' in 2016, built with a chat dialog interface based on IBM's AI platform 'Watson'. The airline has also announced that it is working together with professional services firm Accenture on a new AI voice service that will answer passenger requests at check-in counters.

Singapore Airlines launched its own 'Kris' AI chatbot via its website and Facebook page in 2017 to help passengers with pre-flight-related queries, adding a trial version of Kris on Google Assistant in 2018 that provides information on flight status in response to voice queries.

Malaysia Airlines and travel technology specialist Amadeus launched a chatbot created with Google Cloud's Dialogflow last year. "The chatbot enables customers to search, book, and pay for flights, positioning the carrier to meet future demand and grow revenues from digital channels," explains Tim Synan, Regional Director, Google Cloud in Southeast Asia. "Accessible through Facebook Messenger, their chatbot is accessible 24 hours a day, seven days a week, and uses natural language understanding to mimic human conversation, enabling it to act as a travel buddy for customers."

AVA is a potentially far-reaching initiative for AirAsia, already capable of handling up to 25,000 individual queries per second and designed both to help improve customer experience and generate revenue. In addition to giving information on flights and bookings, the virtual assistant allows passengers to book flights, customise their baggage allowance and even play a trivia game.

Although customer engagement platform Ada Support took just a few weeks to build AirAsia's virtual assistant, it remains under continuous development. Customers can currently access AVA via the airline's website, mobile app and Facebook. Later this year, it will be available via WeChat.

Customers can chat with AVA in English, Bahasa Malaysia, Bahasa Indonesia, Thai, Vietnamese, Chinese (simplified and traditional) and Korean languages. Enquiries that are unable to be resolved by AVA are referred to a live chat support team, which the airline is expanding according to demand.

AVA is expected to help optimise revenue for ancillary services such as booking cargo, paying for excess baggage charges, inflight food and beverages, insurance, seat selection and Wi-Fi services, which last year brought RM2.06 billion (US$238k) in revenue for the airline.

"The most exciting reason why chatbots are now top of mind for big brands is that they can help transform customer service from a cost centre to a profit centre," adds Ada Support's Zive. "By automating cross-sell and up-sell opportunities, chatbots allow you to reorient your customer service resources to drive revenue in a material way."

However, the AVA chatbot is not the only artificial intelligence investment that AirAsia has made. The airline's is committed to transform itself into a digital travel firm and has identified artificial intelligence and machine learning as an opportunity to improve safety, performance and customer experience, partnering with Google for development. Last year the airline launched a new system that makes passengers' faces function as their boarding passes speeding up check-in and airport security clearance with facial recognition, while driving a reduction in boarding process time of 2-4 minutes.

In the future, AirAsia will be able to better predict delays and allowing customers to be better informed of schedule changes and is looking at the potential of real- time sentiment analysis giving the carrier timely information on customer satisfaction.

With such investments becoming commonplace, customer expectations for AI-first services are only likely to increase.

"To scale to hundreds of millions of users, it's important to choose a platform that can integrate with popular conversation platforms like Google Assistant, Slack, Facebook Messenger, and other platforms and devices that their consumers are on," says Synan.

Businesses that invest in AI to enhance their customer experience will also find that they'll have to work hard to create engaging experiences and differentiate their AI services from their competitors. The airline industry is a good example, since airline customers routinely compare services and transactions - such as searching for a flight, redeeming air miles, or pre-booking a special meal - on a like-for-like basis. As airlines vie for leadership in AI services, we can expect chatbots and other virtual assistants to be at the forefront of this competition.

These new AI customer engagement platforms also provide an opportunity for airports. According to the International Air Transport Association, or IATA, the top notifications that passengers want to receive are updates on flight status, baggage information and delivery time, and expected wait-times at airport security: all which could be efficiently disseminated via instant messaging. There are plenty of other opportunities for AI messaging to add value too.

"Beyond the airlines themselves, airports have woken up to the opportunity of AI messaging, with Japan's Narita International Airport, Gatwick in the U.K. and Melbourne Airport in Australia all having implemented messaging services," says Chas Sweeting, CEO of natural language food ordering platform Entrée AI. "Passengers are transient, often pressed for time and therefore the instant nature of messaging and chatbots beats downloading an app, or trying to navigate the web while proceeding to your gate in a busy terminal. Beyond the obvious service use cases, meal ordering for delivery to the gate is now being developed using messaging platforms and AI food recognition."

"Airports are usually functional places which people don't think of as brands, per se - so AI messaging is an opportunity for them to finally have a voice," adds Sweeting.

The writer is an entrepreneur, marketing professional and advisor who has worked across almost every sector of technology. He helps companies, startup ventures and public sector organisations develop marketing strategies, digital initiatives and leverage new marketing technologies.