The idea of having a robot book a hotel or a flight for you is intriguing. A virtual personal assistant can learn your preferences and patterns and book itineraries for you. But how effective can Artificial Intelligence be? Our actions are driven by emotions and we usually let a few other things impact our buying decisions other than price, of course. But will the robots understand these reasons and deliver better results? Let’s find out what the current AI is capable of doing in the travel industry.
AI uses machine learning to analyse data and identify and adapt to patterns in user behaviour. We currently have such AI that recommends music based on user’s taste. Such type of AI is also used to help detect credit card fraud. AI can be used in the travel industry to save millions every year for travellers. However, this will work only when the ‘default settings’ are overridden. For example, when we perform a hotel or flight search, the current systems on most OTAs display the popular hotels first and in a flight search, the flights are arranged from low cost to high cost. This pattern has been working for the OTAs since many years and there is a reason behind it. OTAs have been trying to understand and learn the human psychology and what influences our buying decisions. By displaying popular hotels in the first 1-5 positions in the search results (which are not necessarily cheapest), the cheaper hotels listed after them attract immediate attention and users are more likely to buy them without giving another thought. For example, if a room in a popular hotel costs $500 and an immediate result after that is a room for $300, it seems pretty ‘reasonable’ for the user and he/she is likely to buy it without thinking much. Similarly, the flights are displayed from low to high because that is what is most important for users. People also want to book the cheapest flight but do not necessarily want to stay in the cheapest hotel. Many people prefer staying in a good hotel and investing money in it instead of investing in flights. These patterns have led OTAs to adapt and display results in a specific way because such default settings are not designed to produce least expensive itinerary, instead they are designed to push and pull the levers of human psychology. AI doesn’t have such levers or preset options.
Although AI can do wonders, it also has its limitations. AI cannot always understand the correct context – Why we travel to a specific location? Why do we book hotels close to public transit? What if we want to book a hotel in a particular region (even if it is more expensive and there are cheaper options available)? Why do we usually want the window seat in an airplane? Why do some people always want the aisle seat or the middle seat? AI currently cannot understand the emotions or needs that influence these decisions. AI apps can mimic users’ booking decisions but may not understand the underlying reason for those decisions. Unless we show progress in the technology we use, we will end up spending a lot of time in ‘training’ AI apps and this can include answering lot of preset questions and these apps will take some time before they understand our buying patterns.
What are your thoughts on this? Do you think AI will completely change the way we book our travel? Do you think AI will advance to a point where it will perfectly understand our buying patterns? We would like to hear from you. Please drop your comment in the comments section below and get in touch with us on Facebook, Twitter, Google+ and Linkedin.