Are people ready for Chatbots for their Air Travel?

Hey, I wanna go on a family trip to New York during Christmas

Sure, I can see Emirates and JetBlue flying nonstop to New York. You'd prefer taking off in the morning, right?

Right, how much does the Emirates flight cost?

EK201-Dubai to New York on 24th Dec, for 2 adults and 2 children costs AED 20,020 for Economy return fare...The returning flight is EK202, on 1st Jan, 2019...Do you want me to book it or you wanna look at other dates or flight options?

Go ahead and book it, but make it flexible in case something comes up last minute

All done, here's your e-ticket...

Wouldn't it be nice if you could book a vacation just by typing a few messages on your phone or giving a few commands to the smart speaker? While technically it is possible, we still don't have the perfect virtual travel assistant who knows about our preferences, budgets, and family information and curates the best vacation possible, saving us from the hassles of air travel. It requires deep integration between different systems like social network, airline account, bank account etc., to get the level of fluidity as seen in the above conversation. There's also a definite question about security and privacy, which might impede the progress in this area.

Natural Language Processing (NLP) has come to a point where it is difficult to differentiate Chatbots from humans, not only by the way they converse, but also by their understanding of the context. This makes them an ideal candidate to serve as a customer service representative. There's a range of benefits of using Chatbots from increased efficiency & effectiveness to cost reduction. So, what's stopping companies from going all in, and employing Chatbots as their customer service agents? There's one crucial side of this conversation which might hold some clues, the human side.

H U M A N

However evolved the technology may be, its end consumers are humans (for the most part). So far, we've done a great job of building applications around a human-centric design. But even today, not all of us interface with machines as naturally as intended. Many of us will type "mac screenshot" on google, rather than "How to take a screenshot on a mac". We do this not because of that extra second we save, but because of our awareness of the fact that both queries will generate the same result. We know that it is a machine processing our query, so it won't matter. On the contrary, if it was a human, we'd definitely ask the full question. Chatbots are tricky because of this very reason - they are trying to make interaction with a machine feel more human like. You see, by design NLP aims to be intuitive for an everyday user who doesn't or wouldn't want to know the technical details of the underlying technology and would rather focus on getting things done. The robotic responses or the sheer number of options displayed, might overwhelm customers who are used to interacting with human agents.

Source: IATA Global Passenger Survey, 2014

C O M P A N I E S

Airlines and Travel technology companies have already committed big budgets to do research on Chatbots along with other promising emerging technologies. For them, Chatbots' benefits aren't just limited to cost reduction, it is the aggregation of consumer data that they're after. Think about it - right now flights have to be booked through the airline website or an online travel agency, e-ticket and boarding pass are received over email, flight preferences have to be changed within the airline account, flight status updates are sent via SMS. There's an obvious fragmentation of the information flow between the customers and companies. Chatbots can bring all this information together and manage all the communications with users on one single platform. But they do more than that. If users go for social media Chatbots like Facebook Messenger, the Chatbots can fetch their public profile data, which is a goldmine for the companies.

While it might seem like these companies should rush to implement Chatbots, they aren't going at it with a full swing because they don't exhibit one primary characteristic - emotions. Customer Service Agents represent the face of airlines and act as the primary touch point for customers. That's why so much time and effort is spent to train them so that they can provide a personalized experience to their customers. A Chatbot on the other hand, might go off the rails if a customer throws a complex query at it. This would result in a poor experience for customers and could directly hamper with the airline's brand image.

Source: SITA Passenger IT Trends, 2017

A I R T R A V E L

Air Travel bookings have been traditionally dominated by travel agents. Even today, a significant portion of the bookings made are via these agents or online booking sites like Expedia, MakeMyTrip etc. But in recent years the number of bookings made through direct channels like airline websites, mobile apps etc., have been increasing. Customers are increasingly preferring self-service options for their air travel according to the SITA Passenger IT Trends, 2018. These passengers are also reported to have higher level of satisfaction at each stage of their journey than those who don't use self-service options. The message is clear, airlines and travel technology companies have to engage more with their customers and gather a lot more data, to keep up with their changing needs and behaviours. While Chatbots might seem the perfect solution for maintaining track of end-to-end transactions with the customers, there's still some work that has to be done to get the customers to use it.

While conducting a quantitative research on the consumer's perception of the adoption of AI Chatbots, I found that while people were inclined towards the use of Chatbots, a majority of them weren't really sure that they'd want to use it at its current state. There could be multiple reasons for these indecisive responses: privacy, unintuitive interface, inaccurate results, excessive number of steps, low perceived usefulness. Most of the respondents however feel that this technology holds potential in the future and that there'll be more interesting applications in the coming years.

Sample Chatbot

Results of this study shows that both Perceived Usefulness and Perceived Ease of Use are decisive factors for consumer’s attitude towards using Chatbots. Companies should focus on making the transaction feel more natural and easy to understand and use. While the NLP engines being used today are capable of processing and interpreting various formats and styles of the same information, showing too much information at once is found to leave a negative impression of electronic conversational agents.

Companies can also try to identify smaller transactions which might be primarily information based, for example showing flight status, or baggage allowance etc. These transactions are small and the traditional ways of performing them require navigating through different menu items to retrieve the desired information. These functions could be easily done by AI Chatbots as they are contextually aware and can respond quickly. It is also important for airlines and travel technology companies to deploy Chatbots on platforms that are already popular among users - like Facebook Messenger, WhatsApp etc. It is great to see that efforts are already being made by many industry players to put this technology on a fast-track by experimenting and continuously improving it, to provide a better and smoother journey for us.