2017 is shaping up to be a very significant year in terms of the impact of new technologies in the travel industry. It could quite possibly be the most important year since the arrival of the Netscape internet browser back in December 1994.
Interestingly, Netscape was not the first browser as that title is held by the browser simply called "WorldWideWeb", which was released in 1990 by Sir Tim Berners-Lee and then renamed to Nexus in 1991 before being retired in 1994. Why is this relevant? Well, it is a great example of how technology often follows a pattern where the first product is responsible for opening a new market but then is overtaken by followers who have learned from the challenges and successes of the initial solution.
Furthermore, the internet browser, once the infrastructure was fully ready, became the catalyst for more than 20 years of innovation in the travel industry.
Today's 'buzzword' tech
The term blockchain was first used in 2008 as part of the original bitcoin white paper published by 'Satoshi Nakamoto' (the actual identity is still unknown) and in 2009 the bitcoin digital currency was launched. Once again we saw a significant lack of traction for this new technology and it was only in 2013 the number of active bitcoin wallet holders passed the 1 million mark (13M at time of writing). However, the word blockchain has suddenly become a very popular term in 2017 and we are likely to see the first serious attempt of adoption of blockchain technology in the travel industry before the end of the year.
The term "chatterbot" — now referred to as "chatbot" — was first used in 1994 to describe software capable of having a conversation with a human user. The very first chatbot was called ELIZA and launched in 1966 so from a neutral perspective we have not been very good at adopting and developing this technology. However, based on the level of interest and attention paid to the chatbot term this year, 2017 looks like the breakthrough year for this 50-year-old technology in the travel industry.
Finally, artificial intelligence (AI) was actually created back in 1956 but interest increased and dipped up to when IBM's "Deep Blue" beat Garry Kasparov at chess in 1997. Since then AI investments have been growing every year and AI technology has become an integrated part of our computing world. However, the travel industry once again has been incredibly slow to adapt AI in its generic form as a problem solving simulation engine — until we suddenly started talking about AI-driven chatbots for travel in 2017.
So until now the travel industry has been unable to take advantage of the unlimited number of customer interactions we have processed via the online travel solutions — in fact all we know about a user when they book via a browser-based or mobile app based form is who they were and what they bought (or at least searched for).
Imagine if we knew their emotional state and were able to detect the likely level of intent to buy vs search based on the words used in an actual conversation. This is the new world of opportunities we are opening up when we introduce active use of AI chatbot technologies and if it is combined with blockchain-powered smart contracts then we are suddenly able to imagine and develop an entire new way of distributing and consuming travel.
Initial observations by some of the early movers
There might be other solutions out there with different trends and patterns and I apologise in advance for not being able to include them in my observations.
Reacting to trigger words
The first patterns are emerging as a result of a growing user base and number of conversations.
Firstly, we know that users are including emotional value words when they have a conversation with an AI chatbot. "Thank you" is the 24th most used expression, while "sorry" is number 44 on the list of 50 most used words; despite the fact that the users know they actually are communicating with automated software.
Obviously a number of negatively loaded words are also being used, but when they can be recognised instantly it is possible to transfer an unhappy user to a human agent instantly, rather than continuing the misery in the AI channel.
While we still are in the early days we are actually starting to see patterns of the words used most when an actual purchase is happening vs words used mostly for initial search. Imagine what this knowledge can represent in terms of prioritising a given user or promoting additional product options.
Chat is mobile-only
Another interesting finding from the early AI chatbot days is the fact that users appear to be completely ready to say goodbye to the browser-based forms and use free conversation in a mobile-only user interface.
This is a fascinating development as it actually allows us to provide a personalised style of dialogue with each user followed by delivery of product cards which are designed "just for you". The first result from this intimate conversation is a significantly shortened list of travel options in response to a search query, most users prefer 3-5 options which is where AI really comes into play. The options are selected from the usual hundreds of initial options based on a combination of initial preferences provided, actual behavior, decision making in past conversations and a detailed analysis of the current conversation.
This type of personalisation is simply not possible using the current travel technology infrastructure as the point of sales systems used are purely transactional (search, list and select) and furthermore no human travel consultant has time or ability to look at all past search results, actual bookings and conversations for a given traveller before making a recommendation for the next trip.
Assuming the new AI chatbot technology solutions evolve on a similar curve and speed as the one experienced by browser-based booking engines for the past 20 years, it therefore seems very likely that we are entering a new era of travel technology solutions. One where each traveller is experiencing a fully customisable and personal conversation for each search and booking interaction, and once the new blockchain enabled distribution models emerge we will be able to manage scenarios like the following.
- I am going on my monthly trip to Chicago and simply tell my AI chatbot: "I am going to Chicago again on Tuesday and the second Tuesday in November, please book my travel". A few minutes later my confirmed itinerary shows up
- I am going from San Francisco to London sometime between 15 October and 15 November on a three-day trip and will be prepared to pay 800 USD for my ticket and maximum 200 USD per night for accommodation. Once I have provided my criteria my chosen AI chatbot will query the blockchain world and present the three best options for me based on my past selections. When I select an option the AI chatbot will simply contact the blockchain and execute the relevant smart contract and deliver the details back to me
- I am on the road and need to change my return flight and simply write "please change my return flight to the first flight back to SFO tomorrow". A few minutes later a confirmation appears in my conversation interface after the flight has been changed by a human agent and a hotel booking has been made near the airport by the AI chatbot
The fascinating part of the above scenarios is that none of them require any other user interface than the ability for me to write my request in free text. Then a combination of AI and human driven service will deliver the result back to me; I can be actively involved in other events while the service is happening rather than waiting for my browser page to show a long list of mostly irrelevant options, or even worse — timeout because I got a call in the middle of my search for information.
What's ahead?
As described above, the technical components required to create these solutions actually exist. AI chatbot technology is already here and will be adopted rapidly by millennial users as well as very frequent travellers who simply want the most personal and intelligent travel service. Blockchain is also here but we are in the very early stages of connecting travel inventory to this distribution platform and it will be probably take 1-2 years before we see any meaningful booking volumes from this channel.
Given the global interest in developing innovative travel solutions the biggest obstacle to change (or cause of delay) is likely to be the lack of decision making from existing travel suppliers about how to start supporting the new solutions. However, that will only encourage new players and their investors to move even faster with the product development, so I firmly believe that the AI chatbot channel will represent a significant (20% or more) share of the corporate booking world by 2020.
During the next 3-5 years we will probably see a large number of AI chatbot startups evolve focusing on different user groups and business models before the eventual consolidation happens just like we have seen it happen in the online booking engine space and almost any other area of technology.
There is already a critical shortage of AI software development experts (75,000 openings in the US alone) and this will lead to a high demand for the early solutions as a large number of the traditional travel suppliers and service providers will be unable to develop their own AI chatbot solutions due to a lack of technical expertise.
So fasten your seatbelt and get ready to finally experience a first class service delivered by your own personal travel assistant who will know your habits, preferences and decision making process better than anyone else. Once you start trusting this new service you will eventually stop using any other service as it will be unable to recognise you; the race has started for securing market share in the new AI chatbot space.